[{"id":"jev-ultrafast","name":"jev-ultrafast","author":"browser-use","url":"https://github.com/browser-use/jev-ultrafast","category":"Browser & OS Action","plainSummary":"给浏览器一个目标，让 Jev 选择操作和页面控件，需要输入文字时再调用文本模型。","plainSummaryEn":"A browser Agent that uses Jev to choose actions and page controls, calling a text model only when input is needed.","jevDecisionPoint":"根据当前 DOM，在一次请求里选择操作和相应控件；文本模型负责生成输入内容。","jevDecisionPointEn":"Chooses an action and its matching DOM element in one request; a text model generates input text.","highlightBenefit":"把界面选择、文字生成与浏览器执行分开，方便查看每一步。","highlightBenefitEn":"Separates UI choices, text generation, and browser execution so each step can be inspected.","tags":["browser-automation","typed-decisions"],"stars":15495,"forks":384,"openIssues":36,"license":"MIT","language":"Python","lastCommitAt":"2026-09-18T16:28:35Z","createdAt":"2026-09-16T21:30:12Z","metadataFetchedAt":"2026-09-21T21:53:29.929Z","avatarUrl":"https://avatars.githubusercontent.com/u/192012301?v=4","summarySource":"human-reviewed","claimStatus":"作者在单一 Google Flights 任务上做了每组 3 次对照，报告中位数约 7.1 秒；初始导航未计入，不是通用 Agent 基准。","claimStatusEn":"The author reports a median of about 7.1 seconds on one Google Flights task, with three runs per setup. Initial navigation is excluded; this is not a general Agent benchmark and was not independently retested.","plainSummaryJa":"Jev が操作とページ要素を選び、入力が必要なときだけテキストモデルを呼ぶブラウザー Agent。","plainSummaryKo":"Jev가 동작과 페이지 요소를 선택하고, 입력이 필요할 때만 텍스트 모델을 호출하는 브라우저 Agent.","jevDecisionPointJa":"現在の DOM から操作と対応する要素を一度に選び、入力文は別モデルが生成する。","jevDecisionPointKo":"현재 DOM에서 동작과 해당 요소를 한 번에 선택하며 입력 문장은 별도 모델이 생성한다.","highlightBenefitJa":"画面上の選択、文章生成、実行を分け、各ステップを確認できる。","highlightBenefitKo":"화면 선택, 텍스트 생성, 실행을 분리해 각 단계를 살펴볼 수 있다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"confirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/browser-use/jev-ultrafast/blob/1231850a0bf1a0c0341fe408ef1668dbbfdfac46/jev_ultrafast/model.py","note":"Fixed-commit source reviewed / 已核对固定版本源码"},{"url":"https://github.com/browser-use/jev-ultrafast/blob/1231850a0bf1a0c0341fe408ef1668dbbfdfac46/README.md","note":"Fixed-commit source reviewed / 已核对固定版本源码"},{"url":"https://github.com/browser-use/jev-ultrafast/blob/1231850a0bf1a0c0341fe408ef1668dbbfdfac46/docs/performance.md","note":"Fixed-commit source reviewed / 已核对固定版本源码"}],"pinned":true},{"id":"typesafe-mcp","name":"typesafe-mcp","author":"itsmostafa","url":"https://github.com/itsmostafa/typesafe-mcp","category":"MCP & Integrations","plainSummary":"让 Claude Code、Claude Desktop、Codex 和 Pi 通过 MCP 或扩展向 Jev 提问，获取结构化判断。","plainSummaryEn":"An MCP server that lets Claude Code, Claude Desktop, Codex and Pi ask Jev typed questions.","jevDecisionPoint":"把状态和 Choice、Score、Noul 问题交给 Jev，返回答案与概率。","jevDecisionPointEn":"Submits state and Choice, Score or Noul questions and returns structured answers with probabilities.","highlightBenefit":"让调用方能检查回答，再直接接到自己的分支判断中。","highlightBenefitEn":"Provides decisions that calling code can inspect and branch on.","tags":["mcp-integrations","coding-agents","typed-decisions"],"stars":210,"forks":8,"openIssues":1,"license":"MIT","language":"Go","lastCommitAt":"2026-09-21T18:04:15Z","createdAt":"2026-09-17T05:23:34Z","metadataFetchedAt":"2026-09-21T21:53:32.859Z","avatarUrl":"https://avatars.githubusercontent.com/u/24395434?v=4","summarySource":"human-reviewed","claimStatus":"固定版本 README 与客户端源码可核对接口。","claimStatusEn":"The pinned README and client source document the interface.","plainSummaryJa":"Claude Code、Claude Desktop、Codex、Pi から Jev に質問する MCP サーバー。","plainSummaryKo":"Claude Code, Claude Desktop, Codex, Pi에서 Jev에 질문하는 MCP 서버.","jevDecisionPointJa":"状態と Choice・Score・Noul の質問を送り、構造化された回答と確率を返す。","jevDecisionPointKo":"상태와 Choice·Score·Noul 질문을 보내 구조화된 답과 확률을 받는다.","highlightBenefitJa":"呼び出し側で回答を確認し、処理を分岐できる。","highlightBenefitKo":"호출한 코드에서 답을 확인하고 처리 경로를 나눌 수 있다.","claimStatusJa":"固定バージョンの文書とソースを確認。実行・性能・安全性は当サイトで独立検証していません。","claimStatusKo":"고정 버전 문서와 소스를 확인했습니다. 실행·성능·안전성은 이 사이트에서 독립적으로 검증하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/itsmostafa/typesafe-mcp/blob/0c9f35d9b1859189fc7e7d01947061f311ca6dde/README.md","note":"README.md"},{"url":"https://github.com/itsmostafa/typesafe-mcp/blob/0c9f35d9b1859189fc7e7d01947061f311ca6dde/cmd/evaluate/client.go","note":"cmd/evaluate/client.go"}],"pinned":true,"topics":["mcp","mcp-tools"]},{"id":"jev-mcp","name":"jev-mcp","author":"jkudish","url":"https://github.com/jkudish/jev-mcp","category":"MCP & Integrations","plainSummary":"给 Agent 提供核对引用、筛查内容、查找、重排、分类、比较和提取等八个 MCP 判断工具。","plainSummaryEn":"Eight MCP tools for evidence checks, content screening, search, reranking, classification, comparison and extraction.","jevDecisionPoint":"用 Choice、Noul 等问题评估证据支持、内容风险和候选相关性。","jevDecisionPointEn":"Uses typed questions to assess evidence support, content risk and candidate relevance.","highlightBenefit":"把判断结果和概率交回调用方，由调用方执行阈值与拦截策略。","highlightBenefitEn":"Returns judgments and probabilities; the caller owns thresholds and enforcement.","tags":["mcp-integrations","classification-ranking","search-retrieval"],"stars":230,"forks":10,"openIssues":0,"license":"MIT","language":"TypeScript","lastCommitAt":"2026-09-21T19:12:13Z","createdAt":"2026-09-17T02:25:38Z","metadataFetchedAt":"2026-09-21T21:53:35.858Z","avatarUrl":"https://avatars.githubusercontent.com/u/260253?v=4","summarySource":"human-reviewed","claimStatus":"当前 README 已从三个工具扩展为八个；150–500 ms 为作者报告，概率输出不保证事实或安全判断正确。 已核对固定版本 README 与集成源码。","claimStatusEn":"The current README documents eight tools. The 150–500 ms figure is author-reported; probabilities do not guarantee factual accuracy or safety. README and integration source reviewed at a fixed commit; not independently run or benchmarked by this site.","plainSummaryJa":"引用確認、内容検査、検索、並べ替え、分類、比較、抽出など、8つの MCP 判断ツールを提供する。","plainSummaryKo":"근거 확인, 콘텐츠 검사, 검색, 재정렬, 분류, 비교, 추출 등 8개 MCP 판단 도구를 제공한다.","jevDecisionPointJa":"型付きの質問で、根拠の支持度、内容のリスク、候補の関連度を評価する。","jevDecisionPointKo":"구조화된 질문으로 근거의 뒷받침, 콘텐츠 위험, 후보 관련성을 평가한다.","highlightBenefitJa":"判断と確率を返し、しきい値や制御の適用は呼び出し側に任せる。","highlightBenefitKo":"판단과 확률을 반환하며 임계값과 차단 정책은 호출 측에서 적용한다.","claimStatusJa":"固定版の README と実装を確認済み。このサイトでは独立した実行・性能検証をしていません。","claimStatusKo":"고정 버전 README와 구현을 확인했습니다. 이 사이트에서 독립적인 실행·성능 검증은 하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/jkudish/jev-mcp/blob/89f88b90c3180184d56961160e7178e64c98e468/src/provider.ts#L127","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/jkudish/jev-mcp/blob/89f88b90c3180184d56961160e7178e64c98e468/README.md","note":"固定版本的实现与说明 / Pinned source and documentation"}],"pinned":true},{"id":"semdecide","name":"SemDecide","author":"sharziki","url":"https://github.com/sharziki/semdecide","category":"CLI & Pipelines","plainSummary":"基于 Python 的语义判断 CLI，可给文本或 JSONL 管道做判断、分类、打分和过滤。","plainSummaryEn":"A Python CLI for semantic predicates, routing, scoring and filtering over text or JSONL.","jevDecisionPoint":"将输入交给 Jev，按返回概率与本地阈值生成结果及退出码。","jevDecisionPointEn":"Sends input to Jev and applies local probability thresholds to outputs and exit codes.","highlightBenefit":"把结构化判断接进 Bash 和 CI，保留明确的输出与失败状态。","highlightBenefitEn":"Adds typed decisions to Bash and CI with explicit output and failure states.","tags":["cli-git-gates","classification-ranking","typed-decisions"],"stars":25,"forks":0,"openIssues":0,"license":"MIT","language":"Python","lastCommitAt":"2026-09-16T17:26:42Z","createdAt":"2026-09-16T17:24:18Z","metadataFetchedAt":"2026-09-21T21:53:38.876Z","avatarUrl":"https://avatars.githubusercontent.com/u/97365286?v=4","summarySource":"human-reviewed","claimStatus":"已核对固定版本 README 与调用源码。","claimStatusEn":"Reviewed the fixed-version README and call path.","plainSummaryJa":"テキストや JSONL を判定・分類・採点・フィルタリングする Python CLI。","plainSummaryKo":"텍스트나 JSONL을 판단·분류·채점·필터링하는 Python CLI.","jevDecisionPointJa":"Jev の回答確率とローカル閾値から結果と終了コードを決める。","jevDecisionPointKo":"Jev 응답 확률과 로컬 임곗값으로 결과와 종료 코드를 결정한다.","highlightBenefitJa":"Bash や CI に型付き判断と明確な失敗状態を組み込める。","highlightBenefitKo":"Bash와 CI에 구조화된 판단과 명시적 실패 상태를 연결한다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/sharziki/semdecide/blob/33cf5c03c50e02e59df3f3ea81f0650f6b791545/README.md","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/sharziki/semdecide/blob/33cf5c03c50e02e59df3f3ea81f0650f6b791545/src/reflex_guard/providers/typesafe.py","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/sharziki/semdecide/blob/33cf5c03c50e02e59df3f3ea81f0650f6b791545/src/reflex_guard/policy.py","note":"固定版本说明与源码；本站未执行该项目。"}],"pinned":true,"topics":["ai","ci","cli","jev","semantic","unix"]},{"id":"jev-codex-router","name":"jev-codex-router","author":"0xNatoshi","url":"https://github.com/0xNatoshi/jev-codex-router","category":"Routing & Cost Optimization","plainSummary":"每轮先让 Jev 判断任务类型与难度，再由本地策略为 Codex 选模型、思考深度和速度档。","plainSummaryEn":"Uses Jev to classify each Codex turn, then applies local rules to choose the model, reasoning effort, and speed mode.","jevDecisionPoint":"对当前轮次的任务层级与推理需求分类，本地规则据此选择模型配置。","jevDecisionPointEn":"Classifies task difficulty and reasoning needs before local policy selects a model configuration.","highlightBenefit":"把模型分配规则和判断记录放在本地，便于回看与调整。","highlightBenefitEn":"Keeps routing policy and decision logs local for inspection and tuning.","tags":["llm-routing-cost","coding-agents"],"stars":168,"forks":5,"openIssues":0,"license":"MIT","language":"Python","lastCommitAt":"2026-09-21T21:52:22Z","createdAt":"2026-09-17T04:14:19Z","metadataFetchedAt":"2026-09-21T21:53:41.942Z","avatarUrl":"https://avatars.githubusercontent.com/u/155581707?v=4","summarySource":"human-reviewed","claimStatus":"约 60% 节省来自作者对 237 轮历史会话的标价重算，固定 Token 量且未计跨模型缓存失效；不是实付账单下降。","claimStatusEn":"The roughly 60% saving is the author’s list-price replay estimate over 237 turns with fixed Token counts, excluding cross-model cache invalidation. It is not a reduction in a paid bill and was not independently retested.","plainSummaryJa":"Codex の各ターンを Jev が分類し、ローカルルールでモデル、推論の深さ、速度を選ぶ。","plainSummaryKo":"Codex의 각 턴을 Jev가 분류하고 로컬 규칙으로 모델, 추론 깊이, 속도를 선택한다.","jevDecisionPointJa":"タスクの難易度と推論の必要量を分類し、ローカルポリシーが設定を決める。","jevDecisionPointKo":"작업 난이도와 추론 필요도를 분류하면 로컬 정책이 모델 설정을 결정한다.","highlightBenefitJa":"振り分けルールと判断ログを手元で確認し調整できる。","highlightBenefitKo":"라우팅 규칙과 판단 로그를 로컬에서 검토하고 조정할 수 있다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"confirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/0xNatoshi/jev-codex-router/blob/8292b519659280884627a962c826ac7721136a64/server/jev_server.py","note":"Fixed-commit source reviewed / 已核对固定版本源码"},{"url":"https://github.com/0xNatoshi/jev-codex-router/blob/8292b519659280884627a962c826ac7721136a64/README.md","note":"Fixed-commit source reviewed / 已核对固定版本源码"},{"url":"https://github.com/0xNatoshi/jev-codex-router/blob/8292b519659280884627a962c826ac7721136a64/BACKTEST.md","note":"Fixed-commit source reviewed / 已核对固定版本源码"}],"pinned":true,"topics":["ai","codex","jev","llm","macos","routing","typesafe"]},{"id":"winnow","name":"Winnow","author":"GhalebDweikat","url":"https://github.com/GhalebDweikat/winnow","category":"Context GC & Filter","plainSummary":"给 Claude Code 的长日志装一道筛子。暂时不相关的内容先藏起来，想看时还能完整找回。","plainSummaryEn":"A context filter for Claude Code that hides irrelevant tool-output blocks and keeps the original text available for recall.","jevDecisionPoint":"逐块判断 Read、Bash、Grep 输出是否有用；保留相关或不确定内容，隐藏高置信度无关块。","jevDecisionPointEn":"Jev scores whether Read, Bash and Grep output blocks are needed; local thresholds keep relevant or uncertain blocks.","highlightBenefit":"减少进入上下文的冗余输出，并保留可召回的原文。","highlightBenefitEn":"Reduces visible output while retaining a way to retrieve hidden text.","tags":["context-compaction","coding-agents"],"stars":49,"forks":0,"openIssues":0,"license":"MIT","language":"Python","lastCommitAt":"2026-09-19T01:04:23Z","createdAt":"2026-09-16T03:31:47Z","metadataFetchedAt":"2026-09-21T21:53:44.862Z","avatarUrl":"https://avatars.githubusercontent.com/u/34279235?v=4","summarySource":"human-reviewed","claimStatus":"默认 Jev，也支持 Haiku adapter；过滤收益与误删风险取决于任务和阈值。","claimStatusEn":"Jev is the default judge, with a Haiku adapter available. Benefits and false-removal risks depend on workload and thresholds; not retested here.","plainSummaryJa":"Claude Code の不要なツール出力を隠し、原文を後から呼び戻せるフィルター。","plainSummaryKo":"Claude Code의 불필요한 도구 출력을 숨기고 원문을 다시 불러오는 필터.","jevDecisionPointJa":"Jev が出力の関連性を判定し、ローカルの閾値で必要・不確かな部分を残す。","jevDecisionPointKo":"Jev가 출력의 관련성을 판단하고 로컬 임계값으로 필요하거나 불확실한 부분을 남긴다.","highlightBenefitJa":"表示する情報を絞りつつ、隠した原文を取得できる。","highlightBenefitKo":"보이는 정보를 줄이면서 숨긴 원문을 다시 확인할 수 있다.","claimStatusJa":"固定バージョンの文書とソースを確認。実行・性能・安全性は当サイトで独立検証していません。","claimStatusKo":"고정 버전 문서와 소스를 확인했습니다. 실행·성능·안전성은 이 사이트에서 독립적으로 검증하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/GhalebDweikat/winnow/blob/51d80b945c74c8384bc47fa817179f668289afd8/README.md","note":"README.md"},{"url":"https://github.com/GhalebDweikat/winnow/blob/51d80b945c74c8384bc47fa817179f668289afd8/docs/experiments/jevlike_experiment.py","note":"docs/experiments/jevlike_experiment.py"},{"url":"https://github.com/GhalebDweikat/winnow/blob/51d80b945c74c8384bc47fa817179f668289afd8/sidecar/src/winnow/judge.py","note":"sidecar/src/winnow/judge.py"}],"pinned":true,"topics":["claude-code","claude-code-plugin","context-management","jev","llm-agents","typesafe"]},{"id":"jev-review","name":"Jev Review","author":"devagrawal09","url":"https://github.com/devagrawal09/jev-review","category":"Codebase & Graph Pathfinding","plainSummary":"分阶段检查 Git diff 或整个代码库，在本地面板里展示可复核的审查线索。","plainSummaryEn":"Reviews a Git diff or a codebase in stages and displays review leads in a local dashboard.","jevDecisionPoint":"依次判断风险、文件、证据片段、问题机制和严重程度，再按规则选择审查路径。","jevDecisionPointEn":"Judges risk, files, evidence regions, mechanisms and severity before rule-based reviewer routing.","highlightBenefit":"把审查线索关联到具体代码，便于人工复核。","highlightBenefitEn":"Connects review leads to concrete code for human inspection.","tags":["coding-agents","ast-code-graph","evaluation-benchmarks"],"stars":477,"forks":14,"openIssues":1,"license":"MIT","language":"TypeScript","lastCommitAt":"2026-09-17T00:29:22Z","createdAt":"2026-09-16T21:45:36Z","metadataFetchedAt":"2026-09-21T21:53:47.871Z","avatarUrl":"https://avatars.githubusercontent.com/u/19648701?v=4","summarySource":"human-reviewed","claimStatus":"源码未实现仅高风险调用大模型的分流；审查结果是线索，不是缺陷证明，也无独立算力节省验证。 已核对固定版本 README 与集成源码。","claimStatusEn":"The implementation does not route only high-risk changes to a larger model. Findings are review leads, not proof of defects; compute savings were not established. README and integration source reviewed at a fixed commit; not independently run or benchmarked by this site.","plainSummaryJa":"Git diff またはコード全体を段階的に確認し、ローカル画面にレビューの手掛かりを表示する。","plainSummaryKo":"Git diff 또는 전체 코드베이스를 단계적으로 검토하고 로컬 화면에 검토 단서를 표시한다.","jevDecisionPointJa":"リスク、ファイル、根拠箇所、原因、重大度を判断し、規則に従ってレビュー経路を選ぶ。","jevDecisionPointKo":"위험, 파일, 근거 구간, 원인과 심각도를 판단한 뒤 규칙에 따라 검토 경로를 고른다.","highlightBenefitJa":"手掛かりを具体的なコードと結び付け、人が確認しやすくする。","highlightBenefitKo":"검토 단서를 실제 코드와 연결해 사람이 확인할 수 있게 한다.","claimStatusJa":"固定版の README と実装を確認済み。このサイトでは独立した実行・性能検証をしていません。","claimStatusKo":"고정 버전 README와 구현을 확인했습니다. 이 사이트에서 독립적인 실행·성능 검증은 하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/devagrawal09/jev-review/blob/31f89602797fb7bea007f8a480bf368bf564954e/src/review/judgments.ts#L38","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/devagrawal09/jev-review/blob/31f89602797fb7bea007f8a480bf368bf564954e/README.md","note":"固定版本的实现与说明 / Pinned source and documentation"}],"pinned":true,"topics":["ai","code-review","jev","typesafe-ai","typescript"]},{"id":"blink","name":"Blink","author":"ellipsis-dev","url":"https://github.com/ellipsis-dev/blink","category":"Codebase & Graph Pathfinding","plainSummary":"用自然语言查找文件：多个 walker 沿目录树逐层搜索。","plainSummaryEn":"Finds files from a natural-language query using multiple walkers through the directory tree.","jevDecisionPoint":"Jev 评估文件和目录名的相关概率，程序据此分配 walker。","jevDecisionPointEn":"Jev scores file and directory names; code distributes walkers using those probabilities.","highlightBenefit":"不先建向量索引；结果展示各路径获得的 walker 比例。","highlightBenefitEn":"Searches without a vector index and reports the share of walkers reaching each path.","tags":["ast-code-graph","search-retrieval"],"stars":48,"forks":0,"openIssues":0,"license":null,"language":"TypeScript","lastCommitAt":"2026-09-16T17:00:27Z","createdAt":"2026-09-16T14:30:49Z","metadataFetchedAt":"2026-09-21T21:53:50.888Z","avatarUrl":"https://avatars.githubusercontent.com/u/80834858?v=4","summarySource":"human-reviewed","claimStatus":"已核对固定版本 README 与调用源码。","claimStatusEn":"Reviewed the fixed-version README and call path.","plainSummaryJa":"自然言語の質問から複数の walker でディレクトリ木を探索し、ファイルを探す。","plainSummaryKo":"자연어 질문을 바탕으로 여러 walker가 디렉터리 트리를 탐색해 파일을 찾는다.","jevDecisionPointJa":"Jev が名前の関連確率を評価し、コードが walker を配分する。","jevDecisionPointKo":"Jev가 파일·폴더 이름의 관련 확률을 평가하면 코드가 walker를 배분한다.","highlightBenefitJa":"ベクトル索引なしで探索し、各パスに到達した walker の比率を示す。","highlightBenefitKo":"벡터 인덱스 없이 탐색하며 각 경로에 도달한 walker 비율을 보여 준다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"unconfirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/ellipsis-dev/blink/blob/a621ede75649303a933828c18c27ad800bb43ef0/README.md","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/ellipsis-dev/blink/blob/a621ede75649303a933828c18c27ad800bb43ef0/src/search.ts","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/ellipsis-dev/blink/blob/a621ede75649303a933828c18c27ad800bb43ef0/src/walkers.ts","note":"固定版本说明与源码；本站未执行该项目。"}],"pinned":true},{"id":"neo4jev","name":"neo4jev","author":"jexp","url":"https://github.com/jexp/neo4jev","category":"Codebase & Graph Pathfinding","plainSummary":"在 Neo4j 图谱里逐步找关系，让 Jev 在每个节点选择下一条边。","plainSummaryEn":"Navigates a Neo4j graph one hop at a time, asking Jev which relationship to follow next.","jevDecisionPoint":"Choice 为相邻关系分配概率，Noul 判断是否到达目标，本地 beam search 保留候选路径。","jevDecisionPointEn":"Choice scores neighboring relationships and Noul checks goal completion; local beam search retains candidate paths.","highlightBenefit":"把自然语言目标对应到可以查看的图谱路径。","highlightBenefitEn":"Connects a natural-language goal to inspectable graph paths.","tags":["database-vector","ast-code-graph"],"stars":75,"forks":5,"openIssues":1,"license":"MIT","language":"Jupyter Notebook","lastCommitAt":"2026-09-18T09:25:47Z","createdAt":"2026-09-16T00:07:56Z","metadataFetchedAt":"2026-09-21T21:53:53.930Z","avatarUrl":"https://avatars.githubusercontent.com/u/67427?v=4","summarySource":"human-reviewed","claimStatus":"图导航演示；无 API key 时会明确标注替代答案。未核实大规模图谱上的加速结论。","claimStatusEn":"A graph-navigation demo with explicitly labeled substitute answers when no API key is available. Large-graph speedup claims have not been verified.","plainSummaryJa":"Neo4j のグラフを一段ずつたどり、次に進む関係を Jev に選ばせる。","plainSummaryKo":"Neo4j 그래프를 한 단계씩 탐색하며 다음 관계를 Jev가 선택한다.","jevDecisionPointJa":"Choice で隣接関係を評価し Noul で到達を判定、ローカルの beam search が候補経路を残す。","jevDecisionPointKo":"Choice로 인접 관계를 평가하고 Noul로 목표 도달을 판단하며 로컬 beam search가 후보 경로를 유지한다.","highlightBenefitJa":"自然言語の目標を確認可能なグラフ経路に結び付ける。","highlightBenefitKo":"자연어 목표를 확인 가능한 그래프 경로와 연결한다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"confirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/jexp/neo4jev/blob/d157bbe496eb91813475156942bef1c6badfb342/src/neo4jev/navigator.py","note":"Fixed-commit source reviewed / 已核对固定版本源码"},{"url":"https://github.com/jexp/neo4jev/blob/d157bbe496eb91813475156942bef1c6badfb342/README.md","note":"Fixed-commit source reviewed / 已核对固定版本源码"}],"pinned":true},{"id":"jev-desktop","name":"jev-desktop","author":"lahfir","url":"https://github.com/lahfir/agent-desktop","category":"Browser & OS Action","plainSummary":"把电脑里的按钮和菜单交给 Jev 来选。它读原生无障碍结构，一步步完成桌面操作。","plainSummaryEn":"An optional Jev skill for agent-desktop that chooses controls and actions from native accessibility data.","jevDecisionPoint":"Jev 同时选择动作与目标，并评估置信度和操作风险；本地策略决定是否执行或停止。","jevDecisionPointEn":"Jev selects a target and action and estimates presence and risk; local policy decides whether to execute.","highlightBenefit":"让主 Agent 不必把整棵界面树塞进上下文。","highlightBenefitEn":"Keeps the full UI tree out of the main Agent context while exposing the selected action.","tags":["desktop-os","typed-decisions"],"stars":1424,"forks":86,"openIssues":21,"license":"Apache-2.0","language":"Rust","lastCommitAt":"2026-09-17T05:32:00Z","createdAt":"2026-02-19T18:00:49Z","metadataFetchedAt":"2026-09-21T21:53:56.872Z","avatarUrl":"https://avatars.githubusercontent.com/u/70215676?v=4","summarySource":"human-reviewed","claimStatus":"已核对独立 Jev 脚本；桌面操作及端到端延迟未在本站实测。","claimStatusEn":"The separate Jev scripts were inspected; desktop behavior and end-to-end latency were not tested here.","plainSummaryJa":"agent-desktop のアクセシビリティ情報から操作対象を選ぶ、追加の Jev skill。","plainSummaryKo":"agent-desktop의 접근성 정보에서 조작 대상을 고르는 선택형 Jev skill.","jevDecisionPointJa":"Jev が対象・操作・存在確率・リスクを判定し、ローカル方針が実行を決める。","jevDecisionPointKo":"Jev가 대상·동작·존재 여부·위험을 판단하고 로컬 정책이 실행을 결정한다.","highlightBenefitJa":"主 Agent に画面ツリー全体を渡さず、選んだ操作を確認できる。","highlightBenefitKo":"주 Agent에 전체 UI 트리를 넘기지 않고 선택한 동작을 확인할 수 있다.","claimStatusJa":"固定バージョンの文書とソースを確認。実行・性能・安全性は当サイトで独立検証していません。","claimStatusKo":"고정 버전 문서와 소스를 확인했습니다. 실행·성능·안전성은 이 사이트에서 독립적으로 검증하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/lahfir/agent-desktop/blob/7a8e4a10281c7319733aa200fd79501f34529716/README.md","note":"README.md"},{"url":"https://github.com/lahfir/agent-desktop/blob/7a8e4a10281c7319733aa200fd79501f34529716/scripts/jev/act.mjs","note":"scripts/jev/act.mjs"}],"pinned":true,"topics":["accessibility","accessibility-api","ai-agents","automation","cli","computer-use","computer-use-agent","computer-use-agents","desktop-automation","macos","mcp","rust"]},{"id":"typesafe-ai-playground","name":"TypeSafe AI Playground","author":"markjaquith","url":"https://github.com/markjaquith/typesafe-ai-playground","category":"CLI & Pipelines","plainSummary":"Rust 命令行实验集，可筛查医疗隐私信息、检查代码注释、分析语气并分类行业和职业。","plainSummaryEn":"Rust CLI experiments for health-information screening, comment review, tone analysis and business or occupation classification.","jevDecisionPoint":"把输入文本交给 Jev，返回独立的 Noul 概率、Score 或候选分类。","jevDecisionPointEn":"Sends text to Jev for separate Noul probabilities, scores or category choices.","highlightBenefit":"提供可直接查看结构化结果的终端实验入口。","highlightBenefitEn":"Offers a terminal interface for inspecting structured judgments.","tags":["cli-git-gates","classification-ranking","domain-workflows"],"stars":1,"forks":1,"openIssues":0,"license":"MIT","language":"Rust","lastCommitAt":"2026-09-18T16:38:44Z","createdAt":"2026-09-17T04:01:49Z","metadataFetchedAt":"2026-09-21T21:53:59.921Z","avatarUrl":"https://avatars.githubusercontent.com/u/353790?v=4","summarySource":"human-reviewed","claimStatus":"医疗信息分数仅供筛查，不能单独判断法律意义上的 PHI；原英文 sub-millisecond 无来源，已删除。 已核对固定版本 README 与集成源码。","claimStatusEn":"Health-information scores are screening signals, not a legal PHI determination. The previous sub-millisecond claim had no supporting evidence. README and integration source reviewed at a fixed commit; not independently run or benchmarked by this site.","plainSummaryJa":"医療情報の検査、コメント確認、語調分析、業種・職業分類を試せる Rust CLI。","plainSummaryKo":"의료정보 검사, 주석 검토, 어조 분석, 업종·직업 분류를 실험하는 Rust CLI다.","jevDecisionPointJa":"入力文を Jev に送り、個別の Noul 確率、スコア、分類結果を受け取る。","jevDecisionPointKo":"텍스트를 Jev에 보내 개별 Noul 확률, 점수 또는 분류 결과를 받는다.","highlightBenefitJa":"構造化された判断を端末で確認できる。","highlightBenefitKo":"터미널에서 구조화된 판단 결과를 확인할 수 있다.","claimStatusJa":"固定版の README と実装を確認済み。このサイトでは独立した実行・性能検証をしていません。","claimStatusKo":"고정 버전 README와 구현을 확인했습니다. 이 사이트에서 독립적인 실행·성능 검증은 하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/markjaquith/typesafe-ai-playground/blob/344d77450e1cf3d7aebd2c9566b362b9c693b8e2/src/typesafe.rs#L57","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/markjaquith/typesafe-ai-playground/blob/344d77450e1cf3d7aebd2c9566b362b9c693b8e2/README.md","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/markjaquith/typesafe-ai-playground/blob/344d77450e1cf3d7aebd2c9566b362b9c693b8e2/data/README.md","note":"固定版本的实现与说明 / Pinned source and documentation"}],"pinned":true},{"id":"prism-liquidity-agent","name":"Prism","author":"irfndi","url":"https://github.com/irfndi/prism-liquidity-agent","category":"Domain & Vertical Tools","plainSummary":"观察 Solana 流动性池的 Agent，Jev 提供影子判断与规则结果对照。","plainSummaryEn":"A Solana liquidity Agent with Jev shadow judgments compared against rule-based decisions.","jevDecisionPoint":"评估入池分布、毒性交易流、持有及压力信号，记录用于校准。","jevDecisionPointEn":"Evaluates entry distribution, toxic flow, holding and stress signals for calibration logs.","highlightBenefit":"为确定性交易规则提供可比较的旁路信号，不承诺收益。","highlightBenefitEn":"Adds comparable advisory signals to deterministic trading rules without claiming returns.","tags":["finance-quant","typed-decisions","evaluation-benchmarks"],"stars":66,"forks":5,"openIssues":5,"license":"MIT","language":"TypeScript","lastCommitAt":"2026-09-17T05:13:00Z","createdAt":"2026-06-01T11:29:59Z","metadataFetchedAt":"2026-09-21T21:54:02.863Z","avatarUrl":"https://avatars.githubusercontent.com/u/62242778?v=4","summarySource":"human-reviewed","claimStatus":"已核对固定版本 README 与调用源码。","claimStatusEn":"Reviewed the fixed-version README and call path.","plainSummaryJa":"Solana 流動性プールを監視し、Jev のシャドー判断をルール判断と比較する Agent。","plainSummaryKo":"Solana 유동성 풀을 관찰하고 Jev의 shadow 판단을 규칙 기반 결정과 비교하는 Agent.","jevDecisionPointJa":"入池分布・有害フロー・保有・ストレス信号を評価し、校正用ログに残す。","jevDecisionPointKo":"진입 분포·유해 거래 흐름·보유·시장 스트레스를 평가해 보정용 로그에 기록한다.","highlightBenefitJa":"確定的な取引ルールに比較可能な補助信号を加える。収益は保証しない。","highlightBenefitKo":"결정론적 거래 규칙에 비교할 수 있는 참고 신호를 더하며 수익을 보장하지 않는다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/irfndi/prism-liquidity-agent/blob/22c67bdbe30bab608226832256a5013ad826b707/README.md","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/irfndi/prism-liquidity-agent/blob/22c67bdbe30bab608226832256a5013ad826b707/engine/jev-service.ts","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/irfndi/prism-liquidity-agent/blob/22c67bdbe30bab608226832256a5013ad826b707/engine/jev-gate.ts","note":"固定版本说明与源码；本站未执行该项目。"}],"pinned":true},{"id":"one-v-one-jev","name":"OneVOneJev","author":"emrickgarrett","url":"https://github.com/emrickgarrett/OneVOneJev","category":"High-Frequency & Simulation","plainSummary":"在浏览器里和 Jev 玩 1v1 射击。它读取结构化战况，选择走位、瞄准和开火。","plainSummaryEn":"A browser-based 1v1 shooter where Jev reads structured match state and chooses movement, aim, and firing.","jevDecisionPoint":"每个决策 tick 同时询问移动、视角、开镜、开火与跳跃；API 不可用时使用启发式逻辑。","jevDecisionPointEn":"Each decision tick asks about movement, view, aiming, firing, and jumping, with a heuristic fallback on API failure.","highlightBenefit":"在持续交互的游戏中观察结构化决策。","highlightBenefitEn":"Makes structured decisions visible in an interactive game loop.","tags":["games-simulation","typed-decisions"],"stars":16,"forks":0,"openIssues":0,"license":null,"language":"TypeScript","lastCommitAt":"2026-09-18T03:43:17Z","createdAt":"2026-09-17T01:36:39Z","metadataFetchedAt":"2026-09-21T21:54:05.892Z","avatarUrl":"https://avatars.githubusercontent.com/u/4611485?v=4","summarySource":"human-reviewed","claimStatus":"README 的约 9 Hz 指配置的决策 tick，不是独立测得的持续 API 吞吐量；存在 heuristic fallback。","claimStatusEn":"The README’s roughly 9 Hz refers to the configured decision tick, not independently measured sustained API throughput. A heuristic fallback exists; this site has not retested it.","plainSummaryJa":"構造化された戦況を読み、Jev が移動、照準、射撃を選ぶブラウザー版 1v1 FPS。","plainSummaryKo":"구조화된 전황을 읽고 Jev가 이동, 조준, 발사를 선택하는 브라우저 1v1 FPS.","jevDecisionPointJa":"各 tick で移動、視点、照準、射撃、ジャンプを判断し、API 障害時はヒューリスティックに切り替える。","jevDecisionPointKo":"매 tick마다 이동, 시야, 조준, 발사, 점프를 판단하고 API 오류 시 휴리스틱으로 전환한다.","highlightBenefitJa":"対戦ゲームの中で構造化された判断を観察できる。","highlightBenefitKo":"상호작용하는 게임에서 구조화된 판단을 관찰할 수 있다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"unconfirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/emrickgarrett/OneVOneJev/blob/365b339d04446836352687b3650762106ea37f17/server/src/jev.ts","note":"Fixed-commit source reviewed / 已核对固定版本源码"},{"url":"https://github.com/emrickgarrett/OneVOneJev/blob/365b339d04446836352687b3650762106ea37f17/README.md","note":"Fixed-commit source reviewed / 已核对固定版本源码"}],"pinned":true},{"id":"typesafe-on-neon","name":"safer-with-jev","author":"andrelandgraf","url":"https://github.com/andrelandgraf/safer-with-jev","category":"Security & Guardrails","plainSummary":"给 HTTP 请求装一道内容门禁。Jev 先检查注入指令或不安全内容，通过了再转发到指定地址。","plainSummaryEn":"An HTTP inspection gateway that checks request content with Jev before optionally forwarding it to an HTTPS destination.","jevDecisionPoint":"检查请求内容是否允许通过；本地代码在 pass 时转发，review 或 block 时拦住。","jevDecisionPointEn":"Jev judges prompt injection or unsafe content; local rules forward, flag or block the request.","highlightBenefit":"把内容检查放到上游请求前，直接复用现有 HTTP 服务。","highlightBenefitEn":"Adds an inspectable content check ahead of an existing HTTP service.","tags":["security-guardrails","typed-decisions"],"stars":4,"forks":0,"openIssues":1,"license":null,"language":"TypeScript","lastCommitAt":"2026-09-18T23:50:02Z","createdAt":"2026-09-17T00:01:34Z","metadataFetchedAt":"2026-09-21T21:54:08.881Z","avatarUrl":"https://avatars.githubusercontent.com/u/15163028?v=4","summarySource":"human-reviewed","claimStatus":"仓库已更名为 safer-with-jev；源码提供内容门禁，不证明绝对安全或零延迟。","claimStatusEn":"Renamed to safer-with-jev. The code implements content checks, not a guarantee of safety or zero latency.","plainSummaryJa":"Jev で本文を検査し、条件を満たせば HTTPS 宛先に転送するゲートウェイ。","plainSummaryKo":"Jev로 본문을 검사하고 조건을 충족하면 HTTPS 대상에 전달하는 게이트웨이.","jevDecisionPointJa":"Jev が指示注入や不適切な内容を判定し、ローカル規則が転送・確認・遮断を決める。","jevDecisionPointKo":"Jev가 프롬프트 주입이나 유해 내용을 판단하고 로컬 규칙이 전달·검토·차단을 결정한다.","highlightBenefitJa":"既存 HTTP サービスの前に内容確認を追加できる。","highlightBenefitKo":"기존 HTTP 서비스 앞에 확인 가능한 콘텐츠 검사를 추가한다.","claimStatusJa":"固定バージョンの文書とソースを確認。実行・性能・安全性は当サイトで独立検証していません。","claimStatusKo":"고정 버전 문서와 소스를 확인했습니다. 실행·성능·안전성은 이 사이트에서 독립적으로 검증하지 않았습니다.","licenseStatus":"unconfirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/andrelandgraf/safer-with-jev/blob/b4fb82b129caa46480de6a66387ea2ff053c6511/README.md","note":"README.md"},{"url":"https://github.com/andrelandgraf/safer-with-jev/blob/b4fb82b129caa46480de6a66387ea2ff053c6511/web/src/components/JevIntro.tsx","note":"web/src/components/JevIntro.tsx"},{"url":"https://github.com/andrelandgraf/safer-with-jev/blob/b4fb82b129caa46480de6a66387ea2ff053c6511/src/lib/judge.ts","note":"src/lib/judge.ts"}],"pinned":true},{"id":"tamaratran:fast-jev-compaction","name":"fast-jev-compaction","author":"tamaratran","url":"https://github.com/tamaratran/fast-jev-compaction","category":"Context GC & Filter","plainSummary":"为 Claude Code 删减旧工具调用和结果，保留留下来的原文，不另写摘要。","plainSummaryEn":"Prunes old tool calls and outputs for Claude Code while retaining surviving text verbatim.","jevDecisionPoint":"分别判断工具调用与完整结果是否仍需保留，本地规则执行保留、截短或删除。","jevDecisionPointEn":"Separately judges whether a tool call and its full output are still needed; code keeps, truncates or drops them.","highlightBenefit":"减少摘要改写对路径、命令和报错原文的影响。","highlightBenefitEn":"Avoids rewriting paths, commands and errors into a new summary.","tags":["context-compaction","coding-agents"],"stars":5927,"forks":330,"openIssues":65,"license":"MIT","lastCommitAt":"2026-09-17T22:29:18Z","createdAt":"2026-09-17T05:57:20Z","metadataFetchedAt":"2026-09-21T21:54:11.863Z","avatarUrl":"https://avatars.githubusercontent.com/u/65584429?v=4","summarySource":"human-reviewed","claimStatus":"请求会按预算分批，并非固定一次调用；可能误删上下文，演示动画不是在线 API 性能测试。 已核对固定版本 README 与集成源码。","claimStatusEn":"Requests are split to fit budgets, rather than always using one call. Pruning can remove useful context; the animation is not a live API benchmark. README and integration source reviewed at a fixed commit; not independently run or benchmarked by this site.","plainSummaryJa":"Claude Code の古いツール呼び出しと結果を削減し、残す内容は原文のまま保持する。","plainSummaryKo":"Claude Code의 오래된 도구 호출과 결과를 줄이고 남기는 내용은 원문 그대로 유지한다.","jevDecisionPointJa":"呼び出しと結果全体の必要性を別々に判断し、コードが保持・短縮・削除する。","jevDecisionPointKo":"도구 호출과 전체 결과의 필요성을 각각 판단해 코드가 유지·축약·삭제한다.","highlightBenefitJa":"パス、コマンド、エラーを新しい要約に書き換えずに扱える。","highlightBenefitKo":"경로, 명령, 오류를 새 요약으로 바꾸는 일을 줄인다.","claimStatusJa":"固定版の README と実装を確認済み。このサイトでは独立した実行・性能検証をしていません。","claimStatusKo":"고정 버전 README와 구현을 확인했습니다. 이 사이트에서 독립적인 실행·성능 검증은 하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/tamaratran/fast-jev-compaction/blob/e3f262a7f4d42bd8dd32ced30d26176f7cb545b0/src/client.ts#L1","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/tamaratran/fast-jev-compaction/blob/e3f262a7f4d42bd8dd32ced30d26176f7cb545b0/README.md","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/tamaratran/fast-jev-compaction/blob/e3f262a7f4d42bd8dd32ced30d26176f7cb545b0/hooks/README.md","note":"固定版本的实现与说明 / Pinned source and documentation"}],"pinned":false},{"id":"niazmorshed2007:jev-review","name":"jev-review","author":"NiazMorshed2007","url":"https://github.com/NiazMorshed2007/jev-review","category":"Evaluation & Observability","plainSummary":"供编码 Agent 使用的本地 MCP 代码质量检查器，返回多个维度的结构化评分。","plainSummaryEn":"A local MCP code-quality reviewer returning structured scores to coding Agents.","jevDecisionPoint":"Jev 对正确性、复杂度、测试和安全等维度评分，程序汇总优先改进项。","jevDecisionPointEn":"Jev scores correctness, complexity, tests and security; code ranks areas to improve.","highlightBenefit":"可比较两次检查的评分变化，代码修改仍由主 Agent 负责。","highlightBenefitEn":"Compares scores across checkpoints while leaving code changes to the primary Agent.","tags":["coding-agents","mcp-integrations","evaluation-benchmarks"],"stars":194,"forks":19,"openIssues":5,"license":"MIT","language":"TypeScript","lastCommitAt":"2026-09-17T17:15:18Z","createdAt":"2026-09-17T05:24:51Z","metadataFetchedAt":"2026-09-21T21:54:14.877Z","avatarUrl":"https://avatars.githubusercontent.com/u/77217706?v=4","summarySource":"human-reviewed","claimStatus":"已核对固定版本 README 与调用源码。","claimStatusEn":"Reviewed the fixed-version README and call path.","plainSummaryJa":"コーディング Agent に複数の品質スコアを返すローカル MCP レビューサーバー。","plainSummaryKo":"코딩 Agent에 여러 품질 점수를 반환하는 로컬 MCP 코드 리뷰 서버.","jevDecisionPointJa":"Jev が正しさ・複雑度・テスト・セキュリティを採点し、改善項目を整理する。","jevDecisionPointKo":"Jev가 정확성·복잡도·테스트·보안을 채점하면 코드가 개선 항목의 우선순위를 정한다.","highlightBenefitJa":"チェック間のスコア変化を比較でき、修正は主 Agent が担当する。","highlightBenefitKo":"검사 간 점수 변화를 비교하며 코드 수정은 주 Agent가 담당한다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/NiazMorshed2007/jev-review/blob/57690af54ef7d862c2483342c1e61c14dffcf727/README.md","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/NiazMorshed2007/jev-review/blob/57690af54ef7d862c2483342c1e61c14dffcf727/src/jev/client.ts","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/NiazMorshed2007/jev-review/blob/57690af54ef7d862c2483342c1e61c14dffcf727/src/evaluation/review.ts","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/NiazMorshed2007/jev-review/blob/57690af54ef7d862c2483342c1e61c14dffcf727/src/evaluation/transform.ts","note":"固定版本说明与源码；本站未执行该项目。"}],"pinned":false,"topics":["agent-plugin","ai-agents","claude-code","code-review","codex","coding-agents","cursor","developer-tools","jev","local-first","mcp","model-context-protocol","opencode","software-quality","typescript"]},{"id":"realzachi:pg-jev","name":"pg-jev","author":"realZachi","url":"https://github.com/realZachi/pg-jev","category":"Data & Search","plainSummary":"在 PostgreSQL 查询中用自然语言给数据行筛选、分类和排序。","plainSummaryEn":"Adds natural-language filtering, classification, and ranking of rows to PostgreSQL queries.","jevDecisionPoint":"将行内容交给 Jev，返回是否匹配、类别或等级分数，供 SQL 条件与排序使用。","jevDecisionPointEn":"Sends row content to Jev for a match decision, category, or score used by SQL conditions and ordering.","highlightBenefit":"把语义条件与已有 SQL 查询放在一起，并复用缓存结果。","highlightBenefitEn":"Combines semantic conditions with existing SQL and reuses cached results.","tags":["database-vector","classification-ranking"],"stars":280,"forks":14,"openIssues":1,"license":null,"language":"Shell","lastCommitAt":"2026-09-18T19:23:44Z","createdAt":"2026-09-17T20:04:00Z","metadataFetchedAt":"2026-09-21T21:54:17.866Z","avatarUrl":"https://avatars.githubusercontent.com/u/109552121?v=4","summarySource":"human-reviewed","claimStatus":"数据行会发送到外部 API。作者的小样本耗时和缓存命中耗时不能当作通用模型推理速度。","claimStatusEn":"Row content is sent to an external API. The author’s small-sample timing and cache-hit timing are not general model-inference benchmarks.","plainSummaryJa":"PostgreSQL の行を自然言語で絞り込み、分類、順位付けする。","plainSummaryKo":"PostgreSQL의 행을 자연어로 필터링하고 분류하며 순위를 매긴다.","jevDecisionPointJa":"行の内容を Jev に送り、一致判定、分類、スコアを SQL 条件や並べ替えに使う。","jevDecisionPointKo":"행 내용을 Jev에 보내 일치 여부, 범주, 점수를 SQL 조건과 정렬에 사용한다.","highlightBenefitJa":"既存の SQL に意味的な条件を加え、キャッシュ結果を再利用できる。","highlightBenefitKo":"기존 SQL에 의미 기반 조건을 결합하고 캐시 결과를 재사용한다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"confirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/realZachi/pg-jev/blob/afd11fa856d7a2b831a1bfd8ee7f869ce8efcd62/sql/jev--0.2.0.sql","note":"Fixed-commit source reviewed / 已核对固定版本源码"},{"url":"https://github.com/realZachi/pg-jev/blob/afd11fa856d7a2b831a1bfd8ee7f869ce8efcd62/README.md","note":"Fixed-commit source reviewed / 已核对固定版本源码"},{"url":"https://github.com/realZachi/pg-jev/blob/afd11fa856d7a2b831a1bfd8ee7f869ce8efcd62/LICENSE","note":"Fixed-commit source reviewed / 已核对固定版本源码"}],"pinned":false},{"id":"standardagents:jevpilot","name":"jevpilot","author":"standardagents","url":"https://github.com/standardagents/jevpilot","category":"High-Frequency & Simulation","plainSummary":"在浏览器里开一辆小车，让 Jev 从提前算好的路线和速度里选下一步。","plainSummaryEn":"A browser driving simulator where Jev chooses among locally generated paths and speeds.","jevDecisionPoint":"读取路况、附近车辆和候选轨迹，选择转向与速度；碰撞预测和紧急刹车由代码处理。","jevDecisionPointEn":"Jev reads road and traffic state and selects motion; geometry, collision prediction and braking remain local.","highlightBenefit":"把驾驶判断和物理计算拆开，能直接查看每次选择及其概率。","highlightBenefitEn":"Makes candidate trajectories and decision probabilities visible during a simulation.","tags":["games-simulation","typed-decisions"],"stars":151,"forks":20,"openIssues":0,"license":null,"language":"JavaScript","lastCommitAt":"2026-09-17T21:21:05Z","createdAt":"2026-09-17T15:47:59Z","metadataFetchedAt":"2026-09-21T21:54:20.937Z","avatarUrl":"https://avatars.githubusercontent.com/u/247467195?v=4","summarySource":"human-reviewed","claimStatus":"驾驶模拟 demo；请求频率是作者的调度设置，不是现实自动驾驶能力证明。","claimStatusEn":"A driving simulation demo. Request frequency describes the author’s scheduler, not real-world autonomous-driving capability; not retested here.","plainSummaryJa":"ローカルで作る経路と速度の候補から Jev が選ぶブラウザー運転シミュレーター。","plainSummaryKo":"로컬에서 만든 경로와 속도 후보를 Jev가 고르는 브라우저 운전 시뮬레이터.","jevDecisionPointJa":"道路と交通状況から行動を選び、幾何計算・衝突予測・制動はローカルで行う。","jevDecisionPointKo":"도로와 교통 상태로 동작을 선택하며 기하 계산·충돌 예측·제동은 로컬에서 처리한다.","highlightBenefitJa":"シミュレーション中の候補経路と選択確率を確認できる。","highlightBenefitKo":"시뮬레이션 중 후보 경로와 선택 확률을 확인할 수 있다.","claimStatusJa":"固定バージョンの文書とソースを確認。実行・性能・安全性は当サイトで独立検証していません。","claimStatusKo":"고정 버전 문서와 소스를 확인했습니다. 실행·성능·안전성은 이 사이트에서 독립적으로 검증하지 않았습니다.","licenseStatus":"unconfirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/standardagents/jevpilot/blob/e1beeb13b9a928fb76f167f86af584f4ce9cf180/README.md","note":"README.md"},{"url":"https://github.com/standardagents/jevpilot/blob/e1beeb13b9a928fb76f167f86af584f4ce9cf180/src/jev-request.js","note":"src/jev-request.js"},{"url":"https://github.com/standardagents/jevpilot/blob/e1beeb13b9a928fb76f167f86af584f4ce9cf180/server/jev.js","note":"server/jev.js"}],"pinned":false},{"id":"jkudish:jev-browser","name":"jev-browser","author":"jkudish","url":"https://github.com/jkudish/jev-browser","category":"Browser & OS Action","plainSummary":"给定任务与网址后驱动浏览器，返回最终页面、截图和逐步操作记录。","plainSummaryEn":"Drives a browser from a task and URL, returning the final page, screenshot and action trace.","jevDecisionPoint":"从 DOM 控件选择点击、输入、选择或滚动，并判断目标是否完成、流程是否卡住。","jevDecisionPointEn":"Chooses DOM actions and judges goal completion or stalled progress; a separate model can supply typed text.","highlightBenefit":"把动作建议、实际执行与停止原因留在可检查的轨迹中。","highlightBenefitEn":"Records proposed actions, execution and stop reasons for inspection.","tags":["browser-automation","typed-decisions"],"stars":225,"forks":16,"openIssues":1,"license":"MIT","lastCommitAt":"2026-09-21T19:05:23Z","createdAt":"2026-09-17T19:22:28Z","metadataFetchedAt":"2026-09-21T21:54:23.872Z","avatarUrl":"https://avatars.githubusercontent.com/u/260253?v=4","summarySource":"human-reviewed","claimStatus":"已核对固定版本 README 与集成源码。","claimStatusEn":"README and integration source reviewed at a fixed commit; not independently run or benchmarked by this site.","plainSummaryJa":"タスクと URL を受け取りブラウザを操作し、最終ページ、画像、操作履歴を返す。","plainSummaryKo":"작업과 URL을 받아 브라우저를 조작하고 최종 페이지, 스크린샷, 실행 기록을 반환한다.","jevDecisionPointJa":"DOM の操作候補と完了・停滞を判断する。入力文は別のモデルが補える。","jevDecisionPointKo":"DOM 동작과 완료·정체 여부를 판단하며 입력할 문장은 별도 모델이 만들 수 있다.","highlightBenefitJa":"提案した操作、実行結果、停止理由を記録する。","highlightBenefitKo":"제안한 동작, 실행 결과와 중단 이유를 기록한다.","claimStatusJa":"固定版の README と実装を確認済み。このサイトでは独立した実行・性能検証をしていません。","claimStatusKo":"고정 버전 README와 구현을 확인했습니다. 이 사이트에서 독립적인 실행·성능 검증은 하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/jkudish/jev-browser/blob/8d90c51bedbe7cd07596bfaa532ded019a31d2a8/src/navigate.ts#L1","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/jkudish/jev-browser/blob/8d90c51bedbe7cd07596bfaa532ded019a31d2a8/README.md","note":"固定版本的实现与说明 / Pinned source and documentation"}],"pinned":false},{"id":"y0usaf:pi-jev","name":"pi-jev","author":"y0usaf","url":"https://github.com/y0usaf/pi-jev","category":"Security & Guardrails","plainSummary":"Pi 编码 Agent 扩展：执行前提示工具风险，执行后检查秘密泄露与失败类型。","plainSummaryEn":"A Pi extension that flags tool risks before execution and checks output for secrets and failure types.","jevDecisionPoint":"Jev 给破坏性、越权和输出风险打分，配置决定提示还是请求确认。","jevDecisionPointEn":"Jev scores destructive actions, scope and output risks; configuration chooses warnings or confirmation.","highlightBenefit":"默认只观察；API 出错时放行，不能作为独立的安全边界。","highlightBenefitEn":"Defaults to shadow mode and fails open on API errors; it is not a standalone security boundary.","tags":["security-guardrails","coding-agents"],"stars":133,"forks":8,"openIssues":1,"license":"MIT","language":"TypeScript","lastCommitAt":"2026-09-19T18:59:56Z","createdAt":"2026-09-16T05:57:10Z","metadataFetchedAt":"2026-09-21T21:54:26.865Z","avatarUrl":"https://avatars.githubusercontent.com/u/74448287?v=4","summarySource":"human-reviewed","claimStatus":"默认 shadow 且 API 错误放行；源码检查不代表安全性验证。","claimStatusEn":"Defaults to shadow mode and fails open on API errors; source review is not a security validation.","plainSummaryJa":"Pi のツール実行前にリスクを通知し、実行後に秘密情報や失敗の種類を点検する拡張。","plainSummaryKo":"Pi 도구 실행 전 위험을 알리고 실행 후 비밀 정보와 오류 유형을 확인하는 확장.","jevDecisionPointJa":"Jev のリスク評価を設定に従って警告や確認要求に使う。","jevDecisionPointKo":"Jev 위험 점수를 설정에 따라 경고나 확인 요청에 사용한다.","highlightBenefitJa":"既定は shadow モードで、API エラー時は通過する。独立した安全境界ではない。","highlightBenefitKo":"기본값은 shadow 모드이며 API 오류 시 허용하므로 독립적인 보안 경계가 아니다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/y0usaf/pi-jev/blob/b3478fd4ca1ac8ffcb703f6dc8d6069b555f531e/README.md","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/y0usaf/pi-jev/blob/b3478fd4ca1ac8ffcb703f6dc8d6069b555f531e/src/client.ts","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/y0usaf/pi-jev/blob/b3478fd4ca1ac8ffcb703f6dc8d6069b555f531e/src/gate.ts","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/y0usaf/pi-jev/blob/b3478fd4ca1ac8ffcb703f6dc8d6069b555f531e/src/index.ts","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/y0usaf/pi-jev/blob/b3478fd4ca1ac8ffcb703f6dc8d6069b555f531e/src/output.ts","note":"固定版本说明与源码；本站未执行该项目。"}],"pinned":false,"topics":["guardrails","jev","pi-extension","pi-package","typesafe"]},{"id":"jomatsu:pi-jev-auto-mode","name":"pi-jev-auto-mode","author":"jomatsu","url":"https://github.com/jomatsu/pi-jev-auto-mode","category":"Security & Guardrails","plainSummary":"给 Pi 的命令和文件操作增加规则检查，再由 Jev 评估需要进一步判断的操作。","plainSummaryEn":"Adds rule checks to Pi commands and file operations, then uses Jev to assess cases that need further judgment.","jevDecisionPoint":"本地规则先处理拒绝和允许项，Jev 再检查 bash、write、edit 的授权与风险。","jevDecisionPointEn":"Local rules handle explicit denies and allows before Jev assesses authorization and risk for bash, write, and edit.","highlightBenefit":"保留规则与判断记录，方便调整操作门槛。","highlightBenefitEn":"Keeps policy and decision records available for adjusting operation gates.","tags":["security-guardrails","coding-agents"],"stars":21,"forks":2,"openIssues":1,"license":"MIT","language":"TypeScript","lastCommitAt":"2026-09-17T18:22:49Z","createdAt":"2026-09-17T03:14:28Z","metadataFetchedAt":"2026-09-21T21:54:29.909Z","avatarUrl":"https://avatars.githubusercontent.com/u/50577611?v=4","summarySource":"human-reviewed","claimStatus":"默认不确定时拒绝，但可配置 uncertain=allow；模型判断与阈值不能保证防护效果，本站未验证。","claimStatusEn":"Uncertain cases are denied by default, but uncertain=allow can be configured. Model judgments and thresholds do not guarantee protection; this site has not verified effectiveness.","plainSummaryJa":"Pi のコマンドやファイル操作をルールで確認し、追加判断が必要な操作を Jev が評価する。","plainSummaryKo":"Pi의 명령과 파일 작업을 규칙으로 확인한 뒤 추가 판단이 필요한 작업을 Jev가 평가한다.","jevDecisionPointJa":"明示的な拒否と許可を先に処理し、bash、write、edit の権限とリスクを確認する。","jevDecisionPointKo":"명시적 거부와 허용을 먼저 처리하고 bash, write, edit의 권한과 위험을 평가한다.","highlightBenefitJa":"ルールと判断ログを見ながら操作の許可条件を調整できる。","highlightBenefitKo":"정책과 판단 기록을 보며 작업 허용 기준을 조정할 수 있다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"confirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/jomatsu/pi-jev-auto-mode/blob/06a56043088124ed650471a8589fddd8139708f4/src/jev/transport.ts","note":"Fixed-commit source reviewed / 已核对固定版本源码"},{"url":"https://github.com/jomatsu/pi-jev-auto-mode/blob/06a56043088124ed650471a8589fddd8139708f4/README.md","note":"Fixed-commit source reviewed / 已核对固定版本源码"}],"pinned":false,"topics":["auto-mode","coding-agent","guardrail","jev","permission","pi-coding-agent","pi-extension","pi-package","typesafe"]},{"id":"zadescoxp:jev-trades","name":"Jev-Trades","author":"zadescoxp","url":"https://github.com/zadescoxp/Jev-Trades","category":"Domain & Vertical Tools","plainSummary":"用实时加密货币行情练习模拟交易。Jev 给出交易判断，面板展示虚拟仓位和技术指标，不连真实下单接口。","plainSummaryEn":"A dashboard combining live crypto market data with Jev-guided paper trading; no broker or live-order API is connected.","jevDecisionPoint":"读取已完成的分钟 K 线与技术指标，对启用的资产返回结构化交易判断；Python 按置信度和仓位限制更新模拟账户。","jevDecisionPointEn":"Jev reviews completed minute candles and indicators; Python applies confidence and position limits to a simulated portfolio.","highlightBenefit":"把行情、模型信号和模拟持仓放在一起观察，便于检查决策过程。","highlightBenefitEn":"Shows market inputs, model judgments and paper positions together.","tags":["finance-quant","games-simulation"],"stars":22,"forks":3,"openIssues":3,"license":"Apache-2.0","language":"Python","lastCommitAt":"2026-09-18T05:36:18Z","createdAt":"2026-09-17T18:59:50Z","metadataFetchedAt":"2026-09-21T21:54:32.907Z","avatarUrl":"https://avatars.githubusercontent.com/u/78777405?v=4","summarySource":"human-reviewed","claimStatus":"README 明确为 paper trading，且使用共享内存账户；不是盈利或实盘能力证据。","claimStatusEn":"The README explicitly describes paper trading with shared in-memory portfolio state, not evidence of profitability or live-trading readiness.","plainSummaryJa":"暗号資産の市場データと Jev による模擬取引を表示するダッシュボード。実注文 API は接続しない。","plainSummaryKo":"암호자산 시장 데이터와 Jev 기반 모의거래를 보여 주는 대시보드. 실주문 API는 연결하지 않는다.","jevDecisionPointJa":"Jev が確定済みの分足と指標を判定し、Python が制限に従って模擬口座を更新する。","jevDecisionPointKo":"Jev가 확정된 분봉과 지표를 판단하고 Python이 제한에 따라 모의 계좌를 갱신한다.","highlightBenefitJa":"市場入力・モデル判断・模擬ポジションをまとめて確認できる。","highlightBenefitKo":"시장 입력·모델 판단·모의 포지션을 함께 확인할 수 있다.","claimStatusJa":"固定バージョンの文書とソースを確認。実行・性能・安全性は当サイトで独立検証していません。","claimStatusKo":"고정 버전 문서와 소스를 확인했습니다. 실행·성능·안전성은 이 사이트에서 독립적으로 검증하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/zadescoxp/Jev-Trades/blob/01fb18e4484d626f03345eaff2b93b641677b786/README.md","note":"README.md"},{"url":"https://github.com/zadescoxp/Jev-Trades/blob/01fb18e4484d626f03345eaff2b93b641677b786/app/page.tsx","note":"app/page.tsx"},{"url":"https://github.com/zadescoxp/Jev-Trades/blob/01fb18e4484d626f03345eaff2b93b641677b786/pipeline/paper_trader.py","note":"pipeline/paper_trader.py"}],"pinned":false,"topics":["automated-testing","backtesting","crypto-trading-bot-latest","trading-bot-ai-automated"]},{"id":"dannote:jev","name":"jev","author":"dannote","url":"https://github.com/dannote/jev","category":"SDK & Decision Frameworks","plainSummary":"把 Jev 接成 Elixir/OTP 异步进程，在 GenServer 中用模式匹配处理返回结果。","plainSummaryEn":"Integrates Jev as an asynchronous Elixir/OTP process whose replies are handled with GenServer pattern matching.","jevDecisionPoint":"将应用状态和 Choice、Score、Noul 问题发往 TypeSafe，再以消息返回答案。","jevDecisionPointEn":"Sends state and typed questions to TypeSafe, then delivers answers as messages.","highlightBenefit":"可在现有消息处理和监督机制里组织决策流程。","highlightBenefitEn":"Fits decision calls into existing message-handling and supervision patterns.","tags":["multilanguage-sdk","typed-decisions"],"stars":27,"forks":2,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-21T13:09:43Z","createdAt":"2026-09-17T20:53:28Z","metadataFetchedAt":"2026-09-21T21:54:35.870Z","avatarUrl":"https://avatars.githubusercontent.com/u/2206583?v=4","summarySource":"human-reviewed","claimStatus":"已核对固定版本 README 与集成源码。","claimStatusEn":"README and integration source reviewed at a fixed commit; not independently run or benchmarked by this site.","plainSummaryJa":"Jev を Elixir/OTP の非同期プロセスとして組み込み、GenServer のパターンマッチで応答を処理する。","plainSummaryKo":"Jev를 Elixir/OTP 비동기 프로세스로 연결하고 GenServer 패턴 매칭으로 응답을 처리한다.","jevDecisionPointJa":"状態と型付き質問を TypeSafe に送り、回答をメッセージで届ける。","jevDecisionPointKo":"상태와 구조화된 질문을 TypeSafe에 보내 답을 메시지로 전달한다.","highlightBenefitJa":"既存のメッセージ処理や監視の仕組みに判断を組み込める。","highlightBenefitKo":"기존 메시지 처리와 감독 구조 안에 판단 호출을 넣을 수 있다.","claimStatusJa":"固定版の README と実装を確認済み。このサイトでは独立した実行・性能検証をしていません。","claimStatusKo":"고정 버전 README와 구현을 확인했습니다. 이 사이트에서 독립적인 실행·성능 검증은 하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/dannote/jev/blob/09fbb6cbaf32257924c08ba993ca8631adc16056/lib/jev/http.ex#L3","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/dannote/jev/blob/09fbb6cbaf32257924c08ba993ca8631adc16056/README.md","note":"固定版本的实现与说明 / Pinned source and documentation"}],"pinned":false},{"id":"blakestone-x:jev-mcp","name":"jev-mcp","author":"blakestone-x","url":"https://github.com/blakestone-x/jev-mcp","category":"MCP & Integrations","plainSummary":"把 Jev 的分类、打分、是非判断和候选匹配封装为 MCP 工具。","plainSummaryEn":"Wraps Jev classification, scoring, yes/no judgments and candidate matching as MCP tools.","jevDecisionPoint":"MCP 服务器调用 TypeSafe SDK，返回选项、概率与结构化评分。","jevDecisionPointEn":"The MCP server calls the TypeSafe SDK and returns choices, probabilities and scores.","highlightBenefit":"让支持 MCP 的客户端复用同一套判断接口。","highlightBenefitEn":"Provides a shared decision interface for MCP-compatible clients.","tags":["mcp-integrations","classification-ranking","typed-decisions"],"stars":16,"forks":3,"openIssues":2,"license":"MIT","language":"Python","lastCommitAt":"2026-09-16T20:49:10Z","createdAt":"2026-09-16T17:24:16Z","metadataFetchedAt":"2026-09-21T21:54:38.909Z","avatarUrl":"https://avatars.githubusercontent.com/u/264689723?v=4","summarySource":"human-reviewed","claimStatus":"已核对固定版本 README 与调用源码。","claimStatusEn":"Reviewed the fixed-version README and call path.","plainSummaryJa":"Jev の分類・採点・二択判断・候補照合を MCP ツールとして公開する。","plainSummaryKo":"Jev 분류·채점·예/아니요 판단·후보 매칭을 MCP 도구로 제공한다.","jevDecisionPointJa":"MCP サーバーが TypeSafe SDK を呼び、選択肢・確率・スコアを返す。","jevDecisionPointKo":"MCP 서버가 TypeSafe SDK를 호출해 선택지·확률·점수를 반환한다.","highlightBenefitJa":"MCP 対応クライアントで共通の判断インターフェースを使える。","highlightBenefitKo":"MCP 호환 클라이언트에서 공통 판단 인터페이스를 재사용한다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/blakestone-x/jev-mcp/blob/59289a0b472a3ebc3e9abfebba2d59ec2c881863/README.md","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/blakestone-x/jev-mcp/blob/59289a0b472a3ebc3e9abfebba2d59ec2c881863/jev_mcp/client.py","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/blakestone-x/jev-mcp/blob/59289a0b472a3ebc3e9abfebba2d59ec2c881863/jev_mcp/server.py","note":"固定版本说明与源码；本站未执行该项目。"}],"pinned":false},{"id":"abovecolin:ha-jev","name":"HA-Jev","author":"AboveColin","url":"https://github.com/AboveColin/HA-Jev","category":"Domain & Vertical Tools","plainSummary":"把 Jev 的判断变成 Home Assistant 传感器，例如检查衣服是否洗完后一直没取。","plainSummaryEn":"Turns Jev judgments into Home Assistant sensors, such as checking whether finished laundry has been left unattended.","jevDecisionPoint":"读取选定设备与实体状态，返回概率、选项或分数，再由配置阈值触发自动化。","jevDecisionPointEn":"Reads selected entity states and returns probabilities, choices, or scores that configured thresholds can use in automations.","highlightBenefit":"把自然语言条件接入已有的传感器、通知和场景。","highlightBenefitEn":"Connects natural-language conditions to existing sensors, notifications, and scenes.","tags":["domain-workflows","typed-decisions"],"stars":42,"forks":2,"openIssues":1,"license":"MIT","language":"Python","lastCommitAt":"2026-09-21T19:41:20Z","createdAt":"2026-09-17T11:01:48Z","metadataFetchedAt":"2026-09-21T21:54:41.855Z","avatarUrl":"https://avatars.githubusercontent.com/u/6614011?v=4","summarySource":"human-reviewed","claimStatus":"作者明确不建议用于门锁、加热器或烟雾报警等安全关键控制；延迟与判断准确性未由本站验证。","claimStatusEn":"The author explicitly advises against safety-critical controls such as locks, heaters, or smoke alarms. This site has not verified latency or judgment accuracy.","plainSummaryJa":"洗濯物の取り忘れなどを Jev が判断し、Home Assistant のセンサーとして扱う。","plainSummaryKo":"세탁물 방치 여부 같은 Jev 판단을 Home Assistant 센서로 제공한다.","jevDecisionPointJa":"選択したエンティティの状態から確率、選択肢、スコアを返し、設定した閾値で自動化につなぐ。","jevDecisionPointKo":"선택한 엔티티 상태에서 확률, 선택지, 점수를 반환하고 설정한 임계값으로 자동화에 연결한다.","highlightBenefitJa":"自然言語の条件を既存のセンサーや通知、シーンに組み込める。","highlightBenefitKo":"자연어 조건을 기존 센서, 알림, 장면에 통합할 수 있다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"confirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/AboveColin/HA-Jev/blob/1b48f2fa1e94d076b54082f889f04b0083bcd1a8/custom_components/jev/coordinator.py","note":"Fixed-commit source reviewed / 已核对固定版本源码"},{"url":"https://github.com/AboveColin/HA-Jev/blob/1b48f2fa1e94d076b54082f889f04b0083bcd1a8/README.md","note":"Fixed-commit source reviewed / 已核对固定版本源码"}],"pinned":false,"topics":["ai","assist","conversation","custom-components","decision-model","hacs","home-assistant","home-automation","homeassistant","jev","llm","python","smart-home","typesafe","voice-assistant","voice-control"]},{"id":"abdelstark:jev-benchmarks","name":"jev-benchmarks","author":"AbdelStark","url":"https://github.com/AbdelStark/jev-benchmarks","category":"Evaluation & Observability","plainSummary":"把 Jev 和 GLiNER 放到同一批分类题上，除了答对率，也检查概率靠不靠谱。","plainSummaryEn":"A benchmark comparing Jev and GLiNER on text classification, probability calibration and selective automation.","jevDecisionPoint":"对固定文本和标签集合做分类，记录每个标签的概率、耗时和失败。","jevDecisionPointEn":"Runs the same labeled text tasks through both backends and records probabilities, latency and failures.","highlightBenefit":"能看清模型在哪些任务上适合自动处理，在哪些任务上容易过度自信。","highlightBenefitEn":"Helps examine task-specific accuracy and whether confidence scores support chosen thresholds.","tags":["evaluation-benchmarks","classification-ranking"],"stars":15,"forks":3,"openIssues":4,"license":"Apache-2.0","language":"Python","lastCommitAt":"2026-09-17T07:53:49Z","createdAt":"2026-09-17T07:02:47Z","metadataFetchedAt":"2026-09-21T21:54:44.907Z","avatarUrl":"https://avatars.githubusercontent.com/u/45264458?v=4","summarySource":"human-reviewed","claimStatus":"作者报告 300 样本试验；远程 Jev 与本地 GLiNER 的延迟不能当作统一硬件比较。本站未重跑。","claimStatusEn":"The author reports a 300-example pilot. Hosted Jev and local GLiNER timings are not hardware-normalized; this site did not rerun it.","plainSummaryJa":"文章分類・確率校正・選択的自動化を Jev と GLiNER で比較するベンチマーク。","plainSummaryKo":"텍스트 분류·확률 보정·선별적 자동화를 Jev와 GLiNER로 비교하는 벤치마크.","jevDecisionPointJa":"同じラベル付き課題を実行し、確率・遅延・失敗を記録する。","jevDecisionPointKo":"동일한 라벨 데이터로 확률·지연·실패를 기록한다.","highlightBenefitJa":"課題ごとの精度と、閾値に使う確率の妥当性を調べられる。","highlightBenefitKo":"과제별 정확도와 임계값에 쓰는 확률의 적절성을 살펴볼 수 있다.","claimStatusJa":"固定バージョンの文書とソースを確認。実行・性能・安全性は当サイトで独立検証していません。","claimStatusKo":"고정 버전 문서와 소스를 확인했습니다. 실행·성능·안전성은 이 사이트에서 독립적으로 검증하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/AbdelStark/jev-benchmarks/blob/0d610cc53e79bcbec691312b0c4adb4a0e371642/README.md","note":"README.md"},{"url":"https://github.com/AbdelStark/jev-benchmarks/blob/0d610cc53e79bcbec691312b0c4adb4a0e371642/src/jev_benchmarks/io.py","note":"src/jev_benchmarks/io.py"},{"url":"https://github.com/AbdelStark/jev-benchmarks/blob/0d610cc53e79bcbec691312b0c4adb4a0e371642/src/jev_benchmarks/adapters/jev.py","note":"src/jev_benchmarks/adapters/jev.py"}],"pinned":false,"topics":["benchmarking","calibration","evaluation","machine-learning","reproducibility","selective-classification","zero-shot-classification"]},{"id":"gamesonrblx:jevbridge","name":"Jevbridge","author":"gamesonrblx","url":"https://github.com/gamesonrblx/Jevbridge","category":"MCP & Integrations","plainSummary":"通过 ACP、MCP 和命令行，把 Jev 或普通模型接成同一套结构化判断接口。","plainSummaryEn":"Exposes a shared structured-decision interface for Jev and other models through ACP, MCP and a CLI.","jevDecisionPoint":"将状态和有界问题发送给选定后端，统一解析结果，并支持离线规则后端。","jevDecisionPointEn":"Sends state and bounded questions to the chosen backend, with an offline rule backend also available.","highlightBenefit":"让 Agent 在共同接口下比较或替换判断后端。","highlightBenefitEn":"Lets agents compare or replace decision backends through one interface.","tags":["mcp-integrations","cli-git-gates","typed-decisions"],"stars":37,"forks":3,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-21T09:09:06Z","createdAt":"2026-09-18T02:36:01Z","metadataFetchedAt":"2026-09-21T21:54:50.964Z","avatarUrl":"https://avatars.githubusercontent.com/u/316602654?v=4","summarySource":"human-reviewed","claimStatus":"已核对固定版本 README 与集成源码。","claimStatusEn":"README and integration source reviewed at a fixed commit; not independently run or benchmarked by this site.","plainSummaryJa":"ACP、MCP、CLI を通じ、Jev や他のモデルを共通の判断インターフェースにつなぐ。","plainSummaryKo":"ACP, MCP, CLI를 통해 Jev와 다른 모델을 공통 판단 인터페이스로 연결한다.","jevDecisionPointJa":"状態と限定された質問を選択したバックエンドに送り、オフライン規則も利用できる。","jevDecisionPointKo":"상태와 한정된 질문을 선택한 백엔드에 보내며 오프라인 규칙도 지원한다.","highlightBenefitJa":"Agent が共通の入口で判断バックエンドを比較・交換できる。","highlightBenefitKo":"Agent가 하나의 인터페이스로 판단 백엔드를 비교하거나 바꿀 수 있다.","claimStatusJa":"固定版の README と実装を確認済み。このサイトでは独立した実行・性能検証をしていません。","claimStatusKo":"고정 버전 README와 구현을 확인했습니다. 이 사이트에서 독립적인 실행·성능 검증은 하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/gamesonrblx/Jevbridge/blob/54c5587565cb346788d67805ef2412f46a21bb33/src/client.ts#L1","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/gamesonrblx/Jevbridge/blob/54c5587565cb346788d67805ef2412f46a21bb33/README.md","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/gamesonrblx/Jevbridge/blob/54c5587565cb346788d67805ef2412f46a21bb33/brand/README.md","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/gamesonrblx/Jevbridge/blob/54c5587565cb346788d67805ef2412f46a21bb33/src/types.ts","note":"固定版本的实现与说明 / Pinned source and documentation"}],"pinned":false},{"id":"mizchi:jev-playground","name":"jev-playground","author":"mizchi","url":"https://github.com/mizchi/jev-playground","category":"Evaluation & Observability","plainSummary":"MoonBit 与 TypeScript 的 Jev 实验集，覆盖棋类、浏览器、命令风险和小型语言。","plainSummaryEn":"A MoonBit and TypeScript Jev playground covering games, browsers, command risk and small languages.","jevDecisionPoint":"不同实验把候选动作或判断题交给 Jev，再由对应程序执行或记录。","jevDecisionPointEn":"Experiments send action candidates or typed questions to Jev, then execute or record the answers.","highlightBenefit":"提供源码、实验记录和部分可离线重放的样例，便于比较决策设计。","highlightBenefitEn":"Includes source, experiment notes and some offline replays for comparing decision designs.","tags":["evaluation-benchmarks","games-simulation","browser-automation"],"stars":20,"forks":0,"openIssues":0,"license":null,"language":"TypeScript","lastCommitAt":"2026-09-21T17:00:43Z","createdAt":"2026-09-17T15:49:13Z","metadataFetchedAt":"2026-09-21T21:54:54.017Z","avatarUrl":"https://avatars.githubusercontent.com/u/73962?v=4","summarySource":"human-reviewed","claimStatus":"已核对固定版本 README 与调用源码。","claimStatusEn":"Reviewed the fixed-version README and call path.","plainSummaryJa":"ゲーム・ブラウザー・コマンドリスク・小型言語を扱う MoonBit / TypeScript の Jev 実験集。","plainSummaryKo":"게임·브라우저·명령 위험·소형 언어를 다루는 MoonBit·TypeScript Jev 실험 모음.","jevDecisionPointJa":"候補動作や型付き質問を Jev に送り、各プログラムが回答を実行・記録する。","jevDecisionPointKo":"후보 동작이나 구조화된 질문을 Jev에 보내고 각 프로그램이 응답을 실행·기록한다.","highlightBenefitJa":"ソース、実験記録、一部のオフライン再生例で判断設計を比較できる。","highlightBenefitKo":"소스·실험 기록·일부 오프라인 재생 예제로 판단 설계를 비교할 수 있다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"unconfirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/mizchi/jev-playground/blob/92701bb053ffcec7bb77f23f26b6b1b01ad83f8b/README.md","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/mizchi/jev-playground/blob/92701bb053ffcec7bb77f23f26b6b1b01ad83f8b/lib/client.mbt","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/mizchi/jev-playground/blob/92701bb053ffcec7bb77f23f26b6b1b01ad83f8b/cmd/gomoku/main.mbt","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/mizchi/jev-playground/blob/92701bb053ffcec7bb77f23f26b6b1b01ad83f8b/cmd/jev/main.mbt","note":"固定版本说明与源码；本站未执行该项目。"}],"pinned":false},{"id":"jomatsu:zod-jev","name":"zod-jev","author":"jomatsu","url":"https://github.com/jomatsu/zod-jev","category":"SDK & Decision Frameworks","plainSummary":"为 Zod 校验增加语义规则，例如描述是否匹配或文本是否包含个人信息。","plainSummaryEn":"Adds semantic rules to Zod validation, such as checking whether text matches a description or contains personal information.","jevDecisionPoint":"把同次解析的语义条件合并为 Jev Noul 问题，再按概率返回通过、拒绝或不可用。","jevDecisionPointEn":"Combines semantic conditions from one parse into Jev Noul questions and maps probabilities to validation results.","highlightBenefit":"沿用 Zod 的错误结构，将语义校验问题定位到字段。","highlightBenefitEn":"Uses Zod’s error structure to attach semantic validation issues to fields.","tags":["multilanguage-sdk","security-guardrails","typed-decisions"],"stars":7,"forks":0,"openIssues":0,"license":"MIT","language":"TypeScript","lastCommitAt":"2026-09-17T08:47:40Z","createdAt":"2026-09-17T01:12:53Z","metadataFetchedAt":"2026-09-21T21:54:56.903Z","avatarUrl":"https://avatars.githubusercontent.com/u/50577611?v=4","summarySource":"human-reviewed","claimStatus":"这是 Zod 语义校验库，不是 MCP server。不确定或服务不可用会导致校验失败，但不保证隐私识别或业务判断准确。","claimStatusEn":"This is a semantic validation library for Zod, not an MCP server. Uncertainty or service unavailability fails validation, but privacy detection and business judgments are not guaranteed accurate.","plainSummaryJa":"説明との一致や個人情報の有無など、意味に基づくルールを Zod の検証に加える。","plainSummaryKo":"설명과의 일치나 개인정보 포함 여부 같은 의미 기반 규칙을 Zod 검증에 추가한다.","jevDecisionPointJa":"同じ解析の条件を Jev Noul にまとめ、確率を検証結果に変換する。","jevDecisionPointKo":"한 번의 파싱에서 의미 조건을 Jev Noul 질문으로 묶고 확률을 검증 결과로 변환한다.","highlightBenefitJa":"Zod のエラー形式で、問題のあるフィールドを示せる。","highlightBenefitKo":"Zod 오류 형식으로 문제가 있는 필드를 표시할 수 있다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"confirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/jomatsu/zod-jev/blob/700bd256fe94541a2d21044027cc2dbf5036b396/src/judge.ts","note":"Fixed-commit source reviewed / 已核对固定版本源码"},{"url":"https://github.com/jomatsu/zod-jev/blob/700bd256fe94541a2d21044027cc2dbf5036b396/README.md","note":"Fixed-commit source reviewed / 已核对固定版本源码"}],"pinned":false},{"id":"y0usaf:jev-lm","name":"jev-lm","author":"y0usaf","url":"https://github.com/y0usaf/jev-lm","category":"Evaluation & Observability","plainSummary":"把下一个词当选择题，试着用 Jev 拼出句子；也能让它挑选本地草拟的整段续写。","plainSummaryEn":"A word-level generation experiment that asks Jev to select words or verify locally drafted continuations.","jevDecisionPoint":"Choice 选词，Noul 判断候选续写能否接上以及是否该停止。","jevDecisionPointEn":"Choice selects the next word; Noul evaluates candidate chunks and stopping conditions.","highlightBenefit":"用一套小实验看清决策模型拿来生成文字时会遇到什么问题。","highlightBenefitEn":"Exposes the limitations of using a decision model as a text generator.","tags":["evaluation-benchmarks","typed-decisions"],"stars":5,"forks":1,"openIssues":0,"license":"MIT","language":"TypeScript","lastCommitAt":"2026-09-16T06:36:03Z","createdAt":"2026-09-16T06:00:24Z","metadataFetchedAt":"2026-09-21T21:54:59.917Z","avatarUrl":"https://avatars.githubusercontent.com/u/74448287?v=4","summarySource":"human-reviewed","claimStatus":"研究原型；作者的小词表评测未优于 unigram 基线，不能宣称替代通用语言模型。","claimStatusEn":"A research prototype whose reported small-vocabulary evaluation did not beat a unigram baseline; not a general language-model replacement or independently retested here.","plainSummaryJa":"Jev に単語を選ばせ、ローカルで作った続きの文章を検証する生成実験。","plainSummaryKo":"Jev가 단어를 선택하고 로컬에서 만든 후속 문장을 검증하는 생성 실험.","jevDecisionPointJa":"Choice で次の単語を選び、Noul で続きの候補と終了条件を判定する。","jevDecisionPointKo":"Choice로 다음 단어를 고르고 Noul로 문장 후보와 종료 조건을 판단한다.","highlightBenefitJa":"決定モデルを文章生成に使う際の限界を調べられる。","highlightBenefitKo":"결정 모델을 텍스트 생성에 사용할 때의 한계를 살펴볼 수 있다.","claimStatusJa":"固定バージョンの文書とソースを確認。実行・性能・安全性は当サイトで独立検証していません。","claimStatusKo":"고정 버전 문서와 소스를 확인했습니다. 실행·성능·안전성은 이 사이트에서 독립적으로 검증하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/y0usaf/jev-lm/blob/05e7f03bf447073a0bed3255d4de154f953e533d/README.md","note":"README.md"},{"url":"https://github.com/y0usaf/jev-lm/blob/05e7f03bf447073a0bed3255d4de154f953e533d/python/jev_lm/api.py","note":"python/jev_lm/api.py"}],"pinned":false},{"id":"superagents-lab:jev-search","name":"jev-search","author":"superagents-lab","url":"https://github.com/superagents-lab/jev-search","category":"Data & Search","plainSummary":"用自然语言搜网页：Jev 选择搜索来源和时间范围，再给返回的链接排序。","plainSummaryEn":"Searches the web with Jev choosing sources and time ranges, then ranking returned links.","jevDecisionPoint":"判断查询意图、来源与时间条件，并给各条搜索结果打相关性分数。","jevDecisionPointEn":"Judges search intent, source and time settings, and relevance of individual results.","highlightBenefit":"保留链接、摘要、可调整过滤条件和来源失败提示。","highlightBenefitEn":"Keeps links, snippets, editable filters and source-failure warnings visible.","tags":["search-retrieval","classification-ranking"],"stars":371,"forks":48,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-20T18:37:05Z","createdAt":"2026-09-17T18:05:10Z","metadataFetchedAt":"2026-09-21T21:55:02.878Z","avatarUrl":"https://avatars.githubusercontent.com/u/270369317?v=4","summarySource":"human-reviewed","claimStatus":"已核对固定版本 README 与集成源码。","claimStatusEn":"README and integration source reviewed at a fixed commit; not independently run or benchmarked by this site.","plainSummaryJa":"Jev が検索元と期間を選び、取得したウェブリンクを並べ替える検索ツール。","plainSummaryKo":"Jev가 검색 출처와 기간을 고르고 반환된 웹 링크를 정렬하는 검색 도구다.","jevDecisionPointJa":"検索意図、検索元、期間と、各結果の関連度を判断する。","jevDecisionPointKo":"검색 의도, 출처, 기간과 각 결과의 관련성을 판단한다.","highlightBenefitJa":"リンク、抜粋、変更可能な条件、検索元の失敗を表示する。","highlightBenefitKo":"링크, 발췌, 수정 가능한 조건과 출처 오류를 보여 준다.","claimStatusJa":"固定版の README と実装を確認済み。このサイトでは独立した実行・性能検証をしていません。","claimStatusKo":"고정 버전 README와 구현을 확인했습니다. 이 사이트에서 독립적인 실행·성능 검증은 하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/superagents-lab/jev-search/blob/522868762f0637b20bf533f136e930cceb83b9f3/src/lib/typesafe.ts#L48","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/superagents-lab/jev-search/blob/522868762f0637b20bf533f136e930cceb83b9f3/README.md","note":"固定版本的实现与说明 / Pinned source and documentation"}],"pinned":false},{"id":"vinilana:live-jev","name":"live-jev","author":"vinilana","url":"https://github.com/vinilana/live-jev","category":"High-Frequency & Simulation","plainSummary":"浏览器里的俯视小车模拟器，用 Jev 选择车道与速度，并可与聊天模型对跑。","plainSummaryEn":"A browser-based top-down driving simulator using Jev for lane and speed choices, with an optional chat-model comparison.","jevDecisionPoint":"Jev 批量回答变道、速度、危险程度和行人让行问题，本地规则执行动作。","jevDecisionPointEn":"Jev answers lane, speed, hazard and pedestrian-yield questions; local rules apply the actions.","highlightBenefit":"同一条种子路线可比较两种控制器；紧急制动另有本地逻辑。","highlightBenefitEn":"Runs controllers on the same seeded course, with separate local emergency-braking logic.","tags":["games-simulation","typed-decisions"],"stars":13,"forks":7,"openIssues":1,"license":null,"language":"JavaScript","lastCommitAt":"2026-09-18T13:20:57Z","createdAt":"2026-09-18T00:11:45Z","metadataFetchedAt":"2026-09-21T21:55:05.905Z","avatarUrl":"https://avatars.githubusercontent.com/u/17963870?v=4","summarySource":"human-reviewed","claimStatus":"已核对固定版本 README 与调用源码。","claimStatusEn":"Reviewed the fixed-version README and call path.","plainSummaryJa":"Jev が車線と速度を選ぶブラウザー上の俯瞰型運転シミュレーター。チャットモデルとの比較も可能。","plainSummaryKo":"Jev가 차선과 속도를 선택하는 브라우저 주행 시뮬레이터로 채팅 모델과 비교할 수 있다.","jevDecisionPointJa":"車線・速度・危険度・歩行者優先を Jev が判断し、ローカルルールが動作へ変換する。","jevDecisionPointKo":"Jev가 차선·속도·위험·보행자 양보를 판단하고 로컬 규칙이 동작을 적용한다.","highlightBenefitJa":"同じシードのコースで比較でき、緊急ブレーキには別のローカル処理がある。","highlightBenefitKo":"동일한 시드 코스에서 비교하며 긴급 제동은 별도 로컬 로직을 사용한다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"unconfirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/vinilana/live-jev/blob/cd13ab0a7b58ca9b6749a78aed2a729153fa6910/README.md","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/vinilana/live-jev/blob/cd13ab0a7b58ca9b6749a78aed2a729153fa6910/server.js","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/vinilana/live-jev/blob/cd13ab0a7b58ca9b6749a78aed2a729153fa6910/public/js/brain.js","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/vinilana/live-jev/blob/cd13ab0a7b58ca9b6749a78aed2a729153fa6910/public/js/car.js","note":"固定版本说明与源码；本站未执行该项目。"}],"pinned":false},{"id":"devagrawal09:jev-code","name":"jev-code","author":"devagrawal09","url":"https://github.com/devagrawal09/jev-code","category":"Codebase & Graph Pathfinding","plainSummary":"给编程 Agent 提供代码定位、改动意图检查、测试失败整理和审查意见分流。","plainSummaryEn":"Helps coding Agents locate code, check change intent, triage test failures, and organize review findings.","jevDecisionPoint":"在固定工作流中选择任务，对有限的 diff、代码或日志片段做结构化判断。","jevDecisionPointEn":"Selects a fixed workflow and makes structured judgments over bounded diffs, code, or log excerpts.","highlightBenefit":"返回下一步可检查的线索，同时说明未检查的范围。","highlightBenefitEn":"Returns inspectable leads and states which parts were not checked.","tags":["coding-agents","ast-code-graph","evaluation-benchmarks"],"stars":110,"forks":6,"openIssues":2,"license":"MIT","language":"TypeScript","lastCommitAt":"2026-09-19T21:41:15Z","createdAt":"2026-09-17T15:15:24Z","metadataFetchedAt":"2026-09-21T21:55:11.887Z","avatarUrl":"https://avatars.githubusercontent.com/u/19648701?v=4","summarySource":"human-reviewed","claimStatus":"输出是咨询报告而非通过或批准证明；源码状态与 npm 发布状态可能不同。","claimStatusEn":"Outputs are advisory reports, not pass or approval certificates. Source and npm release status may differ.","plainSummaryJa":"コードの位置特定、変更意図の確認、テスト失敗やレビュー指摘の整理を支援する。","plainSummaryKo":"코드 위치 탐색, 변경 의도 확인, 테스트 실패와 리뷰 의견 정리를 돕는다.","jevDecisionPointJa":"固定ワークフローを選び、範囲を限定した diff、コード、ログを評価する。","jevDecisionPointKo":"고정된 워크플로를 선택하고 제한된 diff, 코드, 로그를 평가한다.","highlightBenefitJa":"確認すべき手掛かりと未確認の範囲を返す。","highlightBenefitKo":"다음에 확인할 단서와 확인하지 않은 범위를 함께 반환한다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"confirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/devagrawal09/jev-code/blob/fd092558ebea389c81d44f9b10e826d9a72afaa3/src/adapters/jev.ts","note":"Fixed-commit source reviewed / 已核对固定版本源码"},{"url":"https://github.com/devagrawal09/jev-code/blob/fd092558ebea389c81d44f9b10e826d9a72afaa3/README.md","note":"Fixed-commit source reviewed / 已核对固定版本源码"}],"pinned":false},{"id":"docxology:daf-jev","name":"daf-jev","author":"docxology","url":"https://github.com/docxology/daf-jev","category":"SDK & Decision Frameworks","plainSummary":"把 Jev 常用零件装成一个 Python 工具箱：提问、批量跑样本、看校准情况，再把结果接到程序或 MCP。","plainSummaryEn":"A Python toolkit for Jev requests, batch evaluation, calibration and MCP access.","jevDecisionPoint":"构造 Noul、Choice、Score 问题，接收概率后由本地函数组合评分、路由及置信度门槛。","jevDecisionPointEn":"Builds Noul, Choice and Score questions, then combines returned probabilities through local functions.","highlightBenefit":"同一套接口覆盖调用、失败记录和评估，便于比较问法与阈值。","highlightBenefitEn":"Reuses request, failure-recording and evaluation interfaces across experiments.","tags":["multilanguage-sdk","evaluation-benchmarks","mcp-integrations"],"stars":5,"forks":1,"openIssues":0,"license":"MIT","language":"Python","lastCommitAt":"2026-09-21T21:14:20Z","createdAt":"2026-09-17T17:13:42Z","metadataFetchedAt":"2026-09-21T21:55:14.909Z","avatarUrl":"https://avatars.githubusercontent.com/u/6911384?v=4","summarySource":"human-reviewed","claimStatus":"源码与文档覆盖客户端、Evaluator 和 MCP 工具；作者的批量性能数据未在本站复现。","claimStatusEn":"Source and docs cover the client, Evaluator and MCP tools; the author’s batch-performance results were not reproduced here.","plainSummaryJa":"Jev の呼び出し・バッチ評価・校正・MCP 接続をまとめた Python ツールキット。","plainSummaryKo":"Jev 호출·배치 평가·보정·MCP 연결을 모은 Python 도구 모음.","jevDecisionPointJa":"Noul・Choice・Score の質問を作り、返された確率をローカル関数で組み合わせる。","jevDecisionPointKo":"Noul·Choice·Score 질문을 만들고 반환 확률을 로컬 함수로 조합한다.","highlightBenefitJa":"呼び出し・失敗記録・評価の仕組みを実験間で再利用できる。","highlightBenefitKo":"호출·실패 기록·평가 인터페이스를 여러 실험에 재사용한다.","claimStatusJa":"固定バージョンの文書とソースを確認。実行・性能・安全性は当サイトで独立検証していません。","claimStatusKo":"고정 버전 문서와 소스를 확인했습니다. 실행·성능·안전성은 이 사이트에서 독립적으로 검증하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/docxology/daf-jev/blob/ff7b28515d7f60c9f7182b128b76b90fbfcba66f/README.md","note":"README.md"},{"url":"https://github.com/docxology/daf-jev/blob/ff7b28515d7f60c9f7182b128b76b90fbfcba66f/src/daf_jev/client.py","note":"src/daf_jev/client.py"}],"pinned":false},{"id":"adhyaay-karnwal:jev-chat","name":"jev-chat","author":"adhyaay-karnwal","url":"https://github.com/adhyaay-karnwal/jev-chat","category":"Evaluation & Observability","plainSummary":"实验性聊天解码器：让 Jev 反复选择词或短语，由代码把它们拼成回答。","plainSummaryEn":"A research chat decoder that repeatedly asks Jev to choose words or phrases and assembles them in code.","jevDecisionPoint":"在有限词表与候选回复中做 Choice，比较逐步解码和整句选择策略。","jevDecisionPointEn":"Compares stepwise Choice decoding with selection from complete candidate replies.","highlightBenefit":"提供解码方法、实验轨迹和失败案例供研究。","highlightBenefitEn":"Provides decoder methods, experiment traces and documented failure cases.","tags":["evaluation-benchmarks","typed-decisions"],"stars":3,"forks":0,"openIssues":1,"license":"MIT","lastCommitAt":"2026-09-17T01:46:51Z","createdAt":"2026-09-17T01:13:08Z","metadataFetchedAt":"2026-09-21T21:55:17.947Z","avatarUrl":"https://avatars.githubusercontent.com/u/182171088?v=4","summarySource":"human-reviewed","claimStatus":"作者明确指出自回归不是 Jev 的原生用途，存在重复和语法错误；不能当成成熟聊天模型。 已核对固定版本 README 与集成源码。","claimStatusEn":"The author states that autoregressive decoding is not Jev’s intended use and documents repetitions and grammatical errors. This is a research decoder. README and integration source reviewed at a fixed commit; not independently run or benchmarked by this site.","plainSummaryJa":"Jev に単語や句を繰り返し選ばせ、コードで回答を組み立てる研究用デコーダー。","plainSummaryKo":"Jev가 단어나 구절을 반복 선택하면 코드가 답변을 조립하는 연구용 디코더다.","jevDecisionPointJa":"段階的な Choice デコードと、完成した候補回答の選択を比較する。","jevDecisionPointKo":"단계별 Choice 디코딩과 완성된 후보 답변 선택을 비교한다.","highlightBenefitJa":"手法、実験履歴、失敗例を研究用に公開している。","highlightBenefitKo":"디코딩 방식, 실험 기록과 실패 사례를 제공한다.","claimStatusJa":"固定版の README と実装を確認済み。このサイトでは独立した実行・性能検証をしていません。","claimStatusKo":"고정 버전 README와 구현을 확인했습니다. 이 사이트에서 독립적인 실행·성능 검증은 하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/adhyaay-karnwal/jev-chat/blob/fb92fd33ddd6603aa544fe4988dd8e918f91d6ca/src/jevchat/client.py#L40","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/adhyaay-karnwal/jev-chat/blob/fb92fd33ddd6603aa544fe4988dd8e918f91d6ca/README.md","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/adhyaay-karnwal/jev-chat/blob/fb92fd33ddd6603aa544fe4988dd8e918f91d6ca/experiments/README.md","note":"固定版本的实现与说明 / Pinned source and documentation"}],"pinned":false},{"id":"tumf:jev-cli","name":"jev-cli","author":"tumf","url":"https://github.com/tumf/jev-cli","category":"CLI & Pipelines","plainSummary":"为 Jev 提供 CLI 和 stdio MCP 入口，输入文本或 JSON，输出结构化判断。","plainSummaryEn":"A CLI and stdio MCP server for typed Jev judgments over text or JSON.","jevDecisionPoint":"向 Jev 提交 noul、choice、score 问题，将结果输出为 JSON 或单个值。","jevDecisionPointEn":"Submits noul, choice and score questions and emits JSON or a primary value.","highlightBenefit":"支持文件和 stdin 输入，可接入 Shell 脚本或 MCP 客户端。","highlightBenefitEn":"Accepts files and stdin for use in shell scripts and MCP clients.","tags":["cli-git-gates","mcp-integrations","typed-decisions"],"stars":9,"forks":0,"openIssues":0,"license":"MIT","language":"Python","lastCommitAt":"2026-09-19T00:47:22Z","createdAt":"2026-09-17T23:45:29Z","metadataFetchedAt":"2026-09-21T21:55:20.897Z","avatarUrl":"https://avatars.githubusercontent.com/u/69994?v=4","summarySource":"human-reviewed","claimStatus":"已核对固定版本 README 与调用源码。","claimStatusEn":"Reviewed the fixed-version README and call path.","plainSummaryJa":"テキストや JSON を Jev で判断する CLI と stdio MCP サーバー。","plainSummaryKo":"텍스트나 JSON을 Jev로 판단하는 CLI 및 stdio MCP 서버.","jevDecisionPointJa":"noul・choice・score を問い、JSON または主値を出力する。","jevDecisionPointKo":"noul·choice·score 질문을 보내 JSON이나 주요 값을 출력한다.","highlightBenefitJa":"ファイルと stdin に対応し、Shell スクリプトや MCP クライアントに接続できる。","highlightBenefitKo":"파일과 stdin을 지원해 Shell 스크립트와 MCP 클라이언트에서 사용할 수 있다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/tumf/jev-cli/blob/60441de4f96870686a5388bd6ed714f36eed646b/README.md","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/tumf/jev-cli/blob/60441de4f96870686a5388bd6ed714f36eed646b/src/jev_cli/__init__.py","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/tumf/jev-cli/blob/60441de4f96870686a5388bd6ed714f36eed646b/src/jev_cli/mcp_server.py","note":"固定版本说明与源码；本站未执行该项目。"}],"pinned":false,"topics":["ai","cli","jev","python","typesafe"]},{"id":"ranjan2829:askjev","name":"AskJev","author":"ranjan2829","url":"https://github.com/ranjan2829/AskJev","category":"Browser & OS Action","plainSummary":"通过 MCP 把 Agent 接到浏览器，由 Jev 选择网页动作，并对付款、删除等操作设置确认环节。","plainSummaryEn":"Connects Agents to a browser over MCP, using Jev to choose page actions with confirmation gates for actions such as payment or deletion.","jevDecisionPoint":"从当前网页控件中选动作，并评估操作风险与不可逆程度。","jevDecisionPointEn":"Selects actions from current page controls and assesses risk and irreversibility.","highlightBenefit":"把自动操作与用户确认放在同一工作流中。","highlightBenefitEn":"Combines automated page actions with user confirmation in one workflow.","tags":["browser-automation","mcp-integrations","security-guardrails"],"stars":8,"forks":3,"openIssues":0,"license":"MIT","language":"TypeScript","lastCommitAt":"2026-09-18T22:53:35Z","createdAt":"2026-09-17T22:40:34Z","metadataFetchedAt":"2026-09-21T21:55:23.894Z","avatarUrl":"https://avatars.githubusercontent.com/u/73215366?v=4","summarySource":"human-reviewed","claimStatus":"确认逻辑只能视作附加检查，不能保证阻止所有危险操作；本站未做浏览器实测。","claimStatusEn":"Confirmation logic is an additional check, not a guarantee against all dangerous actions. This site has not tested it in a browser.","plainSummaryJa":"MCP で Agent とブラウザーを接続し、Jev がページ操作を選ぶ。支払いや削除などには確認を挟む。","plainSummaryKo":"MCP로 Agent와 브라우저를 연결하고 Jev가 페이지 동작을 선택하며 결제나 삭제 등에는 확인 단계를 둔다.","jevDecisionPointJa":"現在のページ要素から操作を選び、リスクと取り消し可能性を評価する。","jevDecisionPointKo":"현재 페이지 요소에서 동작을 선택하고 위험과 되돌릴 수 있는 정도를 평가한다.","highlightBenefitJa":"自動操作とユーザー確認を同じ手順にまとめる。","highlightBenefitKo":"자동 작업과 사용자 확인을 하나의 흐름으로 연결한다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"confirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/ranjan2829/AskJev/blob/7beec5eba7c4a294d21013c158cd884e1c24f08b/mcp/src/jev-client.ts","note":"Fixed-commit source reviewed / 已核对固定版本源码"},{"url":"https://github.com/ranjan2829/AskJev/blob/7beec5eba7c4a294d21013c158cd884e1c24f08b/README.md","note":"Fixed-commit source reviewed / 已核对固定版本源码"}],"pinned":false,"topics":["browser-automation","chrome-extension","jev","mv3","typesafe","typescript"]},{"id":"foadsf:jev-for-engineers","name":"jev-for-engineers","author":"Foadsf","url":"https://github.com/Foadsf/jev-for-engineers","category":"Domain & Vertical Tools","plainSummary":"机械与电气工程的八组 Jev 小实验：分派设计任务、检查仿真日志、匹配零件，再由 Python 规则决定怎么处理。","plainSummaryEn":"Eight mechanical and electrical engineering experiments using Jev for task routing, log checks and component selection.","jevDecisionPoint":"对工程文本做分类、候选选择与风险判断；尺寸计算和最终处置仍交给普通代码。","jevDecisionPointEn":"Jev classifies engineering text and candidates; calculations and final actions stay in Python rules.","highlightBenefit":"用合成样例看清工程判断如何接入程序；示例阈值需要用自己的数据重测。","highlightBenefitEn":"Shows how to connect judgments to engineering workflows using synthetic examples.","tags":["domain-workflows","classification-ranking"],"stars":2,"forks":0,"openIssues":1,"license":"MIT","language":"Python","lastCommitAt":"2026-09-16T23:19:13Z","createdAt":"2026-09-16T23:17:23Z","metadataFetchedAt":"2026-09-21T21:55:26.943Z","avatarUrl":"https://avatars.githubusercontent.com/u/12762442?v=4","summarySource":"human-reviewed","claimStatus":"输入为合成样例，不能据此校准生产阈值或证明工程安全。","claimStatusEn":"Inputs are synthetic and do not establish production thresholds or engineering safety.","plainSummaryJa":"タスク振り分け・ログ確認・部品選択を試す、機械・電気工学向けの八つの Jev 実験。","plainSummaryKo":"작업 배분·로그 검사·부품 선택을 다루는 기계·전기공학용 Jev 실험 8개.","jevDecisionPointJa":"Jev が工学的な文章や候補を分類し、計算と最終処理は Python が担う。","jevDecisionPointKo":"Jev가 공학 텍스트와 후보를 분류하고 계산과 최종 처리는 Python이 맡는다.","highlightBenefitJa":"合成例で工学ワークフローへの判断の組み込み方を確認できる。","highlightBenefitKo":"합성 예제로 공학 작업에 판단을 연결하는 방법을 확인한다.","claimStatusJa":"固定バージョンの文書とソースを確認。実行・性能・安全性は当サイトで独立検証していません。","claimStatusKo":"고정 버전 문서와 소스를 확인했습니다. 실행·성능·안전성은 이 사이트에서 독립적으로 검증하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/Foadsf/jev-for-engineers/blob/181e75208ac3b982c80889267a28a1ed718452a9/README.md","note":"README.md"},{"url":"https://github.com/Foadsf/jev-for-engineers/blob/181e75208ac3b982c80889267a28a1ed718452a9/jev.py","note":"jev.py"}],"pinned":false,"topics":["cad","cae","calibrated-confidence","classification","jev","llm","mechanical-engineering","typesafe"]},{"id":"haseeb-heaven:jev-system-one","name":"jev-system-one","author":"haseeb-heaven","url":"https://github.com/haseeb-heaven/jev-system-one","category":"Voice & Conversation","plainSummary":"终端问答界面由 OpenAI 写回答，Jev 决定回答方式、检查草稿并判断是否重写。","plainSummaryEn":"A terminal Q&A app where OpenAI writes answers and Jev sets response policy and reviews drafts.","jevDecisionPoint":"判断回答模式、深度、不确定性和草稿质量，再把结果交给 LangGraph 流程。","jevDecisionPointEn":"Judges response mode, depth, uncertainty and draft quality inside a LangGraph workflow.","highlightBenefit":"可同时查看最终回答与结构化决策报告。","highlightBenefitEn":"Displays the answer alongside a structured decision report.","tags":["voice-conversation","evaluation-benchmarks","typed-decisions"],"stars":4,"forks":0,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-17T02:45:20Z","createdAt":"2026-09-17T02:09:04Z","metadataFetchedAt":"2026-09-21T21:55:29.921Z","avatarUrl":"https://avatars.githubusercontent.com/u/11544739?v=4","summarySource":"human-reviewed","claimStatus":"已核对固定版本 README 与集成源码。","claimStatusEn":"README and integration source reviewed at a fixed commit; not independently run or benchmarked by this site.","plainSummaryJa":"OpenAI が回答を書き、Jev が方針決定と草稿確認を行う端末の質問応答アプリ。","plainSummaryKo":"OpenAI가 답을 작성하고 Jev가 응답 방침과 초안을 검토하는 터미널 질의응답 앱이다.","jevDecisionPointJa":"LangGraph 内で回答形式、詳しさ、不確実性、草稿品質を判断する。","jevDecisionPointKo":"LangGraph 흐름에서 답변 형식, 깊이, 불확실성과 초안 품질을 판단한다.","highlightBenefitJa":"回答と構造化された判断レポートを並べて確認できる。","highlightBenefitKo":"답변과 구조화된 판단 보고서를 함께 볼 수 있다.","claimStatusJa":"固定版の README と実装を確認済み。このサイトでは独立した実行・性能検証をしていません。","claimStatusKo":"고정 버전 README와 구현을 확인했습니다. 이 사이트에서 독립적인 실행·성능 검증은 하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/haseeb-heaven/jev-system-one/blob/66bbfb66556e6daa0f091a04b4e33c3a314eec28/src/jev_system_one/jev.py#L164","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/haseeb-heaven/jev-system-one/blob/66bbfb66556e6daa0f091a04b4e33c3a314eec28/README.md","note":"固定版本的实现与说明 / Pinned source and documentation"}],"pinned":false},{"id":"leftspace89:jevbird","name":"JevBird","author":"leftspace89","url":"https://github.com/leftspace89/JevBird","category":"High-Frequency & Simulation","plainSummary":"让 Jev 玩 Python 版 Flappy Bird：程序先模拟路线，再让模型选择。","plainSummaryEn":"A Python Flappy Bird game where code simulates candidate routes and Jev picks one.","jevDecisionPoint":"每遇到新的管道，Jev 从候选路线中选一条，游戏执行该路线的拍翅计划。","jevDecisionPointEn":"For each new pipe, Jev selects a route and the game executes its scheduled flaps.","highlightBenefit":"可在画面中查看候选轨迹、概率和最终选择。","highlightBenefitEn":"Visualizes candidate trajectories, probabilities and the selected path.","tags":["games-simulation","typed-decisions"],"stars":5,"forks":1,"openIssues":0,"license":"MIT","language":"Python","lastCommitAt":"2026-09-17T22:18:10Z","createdAt":"2026-09-17T22:25:08Z","metadataFetchedAt":"2026-09-21T21:55:32.916Z","avatarUrl":"https://avatars.githubusercontent.com/u/13282563?v=4","summarySource":"human-reviewed","claimStatus":"已核对固定版本 README 与调用源码。","claimStatusEn":"Reviewed the fixed-version README and call path.","plainSummaryJa":"コードで候補軌道をシミュレートし、Jev が選ぶ Python 版 Flappy Bird。","plainSummaryKo":"코드가 후보 궤적을 시뮬레이션하고 Jev가 하나를 고르는 Python Flappy Bird.","jevDecisionPointJa":"新しいパイプごとに経路を選び、ゲームが予定された羽ばたきを実行する。","jevDecisionPointKo":"새 파이프마다 경로를 선택하면 게임이 예정된 날갯짓을 실행한다.","highlightBenefitJa":"候補軌道・確率・選択結果を画面で確認できる。","highlightBenefitKo":"후보 궤적·확률·선택 결과를 화면으로 확인할 수 있다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/leftspace89/JevBird/blob/1e5fd1397382be91ce74ed0ab0cba86027c3470a/README.md","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/leftspace89/JevBird/blob/1e5fd1397382be91ce74ed0ab0cba86027c3470a/jev_pilot.py","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/leftspace89/JevBird/blob/1e5fd1397382be91ce74ed0ab0cba86027c3470a/flappy_bird.py","note":"固定版本说明与源码；本站未执行该项目。"}],"pinned":false},{"id":"hyunjunjeon:jev-judgment","name":"jev-judgment","author":"HyunjunJeon","url":"https://github.com/HyunjunJeon/jev-judgment","category":"Security & Guardrails","plainSummary":"给编程 Agent 增加授权、操作风险和失败原因的判断检查。","plainSummaryEn":"Adds judgment checks for authorization, operation risk, and failure causes to coding Agents.","jevDecisionPoint":"依据对话与命令结果，判断是否需要询问用户、操作是否越界、失败能否重试。","jevDecisionPointEn":"Uses conversation and command results to assess whether to ask the user, whether an action exceeds scope, and whether to retry.","highlightBenefit":"把需要停下确认的节点显式记录下来。","highlightBenefitEn":"Makes checkpoints that may require user confirmation explicit.","tags":["security-guardrails","coding-agents"],"stars":3,"forks":1,"openIssues":1,"license":"MIT","language":"Python","lastCommitAt":"2026-09-17T03:54:07Z","createdAt":"2026-09-17T03:54:12Z","metadataFetchedAt":"2026-09-21T21:55:35.930Z","avatarUrl":"https://avatars.githubusercontent.com/u/15343250?v=4","summarySource":"human-reviewed","claimStatus":"模型概率不能替代用户授权；缺 key 或网络失败时可继续原流程，本站未验证防护效果。","claimStatusEn":"Model probabilities do not replace user authorization. Missing keys or network failures may leave the original workflow running.","plainSummaryJa":"コーディング Agent に権限、操作リスク、失敗原因の判断を追加する。","plainSummaryKo":"코딩 Agent에 권한, 작업 위험, 실패 원인 판단을 추가한다.","jevDecisionPointJa":"会話と実行結果から、ユーザー確認の必要性、権限範囲、再試行の可否を評価する。","jevDecisionPointKo":"대화와 실행 결과로 사용자 확인 필요성, 권한 범위, 재시도 여부를 평가한다.","highlightBenefitJa":"立ち止まって確認すべき箇所を明示できる。","highlightBenefitKo":"멈추고 확인해야 할 지점을 명시할 수 있다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"confirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/HyunjunJeon/jev-judgment/blob/f6056e7cd31d467d6773a0737731c85939499e5e/skills/jev-judgment/scripts/jev.py","note":"Fixed-commit source reviewed / 已核对固定版本源码"},{"url":"https://github.com/HyunjunJeon/jev-judgment/blob/f6056e7cd31d467d6773a0737731c85939499e5e/README.md","note":"Fixed-commit source reviewed / 已核对固定版本源码"}],"pinned":false},{"id":"jtsang4:jev-cli","name":"jev-cli","author":"jtsang4","url":"https://github.com/jtsang4/jev-cli","category":"CLI & Pipelines","plainSummary":"在终端里问 Jev 判断题。输入文本或 JSON，再给出分类、是非或评分问题，拿回脚本能直接读取的 JSON。","plainSummaryEn":"A CLI for asking Jev classification, yes/no and scoring questions over text or JSON input.","jevDecisionPoint":"针对同一份输入做分类、真假判断和分级评分，返回选项及其概率。","jevDecisionPointEn":"Evaluates typed questions over one input and returns choices and probabilities as JSON.","highlightBenefit":"能接收标准输入，把语义校验接进已有脚本和 CI；支持直连 TypeSafe 或走 Vercel 网关。","highlightBenefitEn":"Accepts standard input and supports direct TypeSafe calls or a Vercel gateway.","tags":["cli-git-gates","classification-ranking","typed-decisions"],"stars":2,"forks":0,"openIssues":1,"license":"MIT","language":"TypeScript","lastCommitAt":"2026-09-18T04:02:41Z","createdAt":"2026-09-17T12:36:40Z","metadataFetchedAt":"2026-09-21T21:55:38.938Z","avatarUrl":"https://avatars.githubusercontent.com/u/17686135?v=4","summarySource":"human-reviewed","claimStatus":"已核对 CLI 与 provider 接口；本站未测量延迟、准确率或费用。","claimStatusEn":"The CLI and provider interfaces were inspected; latency, accuracy and cost were not measured here.","plainSummaryJa":"文章や JSON を入力し、Jev に分類・Yes/No・評価を質問する CLI。","plainSummaryKo":"텍스트나 JSON으로 Jev에 분류·예/아니요·평가 질문을 보내는 CLI.","jevDecisionPointJa":"一つの入力に型付き質問を適用し、選択結果と確率を JSON で返す。","jevDecisionPointKo":"하나의 입력에 타입이 지정된 질문을 적용해 선택 결과와 확률을 JSON으로 받는다.","highlightBenefitJa":"標準入力を受け取り、TypeSafe 直結と Vercel gateway に対応する。","highlightBenefitKo":"표준 입력을 지원하며 TypeSafe 직접 호출이나 Vercel gateway를 쓸 수 있다.","claimStatusJa":"固定バージョンの文書とソースを確認。実行・性能・安全性は当サイトで独立検証していません。","claimStatusKo":"고정 버전 문서와 소스를 확인했습니다. 실행·성능·안전성은 이 사이트에서 독립적으로 검증하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/jtsang4/jev-cli/blob/6ea8abdf702d2fc28bfffd979e83db8c32f3ceb5/README.md","note":"README.md"},{"url":"https://github.com/jtsang4/jev-cli/blob/6ea8abdf702d2fc28bfffd979e83db8c32f3ceb5/src/providers/jev.ts","note":"src/providers/jev.ts"}],"pinned":false,"topics":["cli","jev","skills","typesafe"]},{"id":"buchmark:claude-jev","name":"claude-jev","author":"buchmark","url":"https://github.com/buchmark/claude-jev","category":"Codebase & Graph Pathfinding","plainSummary":"给 Claude Code 的审查发现、排错假设、设计方案和代码搜索结果增加一次 Jev 复核。","plainSummaryEn":"Adds Jev checks to Claude Code review findings, debugging hypotheses, design options and search results.","jevDecisionPoint":"对候选缺陷、解释或选项按预设问题打分，再由阈值与本地策略处理。","jevDecisionPointEn":"Scores candidate defects, explanations or options against preset questions and local policies.","highlightBenefit":"把第二次判断及概率保留下来，便于检查分歧。","highlightBenefitEn":"Keeps a second judgment and its probabilities available for comparison.","tags":["coding-agents","ast-code-graph","evaluation-benchmarks"],"stars":5,"forks":0,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-18T03:33:10Z","createdAt":"2026-09-18T03:27:22Z","metadataFetchedAt":"2026-09-21T21:55:41.913Z","avatarUrl":"https://avatars.githubusercontent.com/u/73613787?v=4","summarySource":"human-reviewed","claimStatus":"已核对固定版本 README 与集成源码。","claimStatusEn":"README and integration source reviewed at a fixed commit; not independently run or benchmarked by this site.","plainSummaryJa":"Claude Code の指摘、原因仮説、設計案、検索結果に Jev の確認を追加する。","plainSummaryKo":"Claude Code의 검토 결과, 원인 가설, 설계안과 검색 결과에 Jev 검사를 추가한다.","jevDecisionPointJa":"候補の問題や選択肢を定義済みの質問で評価し、ローカル規則で処理する。","jevDecisionPointKo":"후보 문제나 선택지를 미리 정한 질문으로 평가하고 로컬 정책으로 처리한다.","highlightBenefitJa":"追加判断と確率を残し、判断の違いを確認できる。","highlightBenefitKo":"추가 판단과 확률을 남겨 의견 차이를 확인할 수 있다.","claimStatusJa":"固定版の README と実装を確認済み。このサイトでは独立した実行・性能検証をしていません。","claimStatusKo":"고정 버전 README와 구현을 확인했습니다. 이 사이트에서 독립적인 실행·성능 검증은 하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/buchmark/claude-jev/blob/06fe407f2894bc5c44f2dfc41e05d3592e5215a5/src/infrastructure/typesafe-jev.ts#L114","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/buchmark/claude-jev/blob/06fe407f2894bc5c44f2dfc41e05d3592e5215a5/README.md","note":"固定版本的实现与说明 / Pinned source and documentation"}],"pinned":false},{"id":"tontoko:jev-browser","name":"jev-browser","author":"tontoko","url":"https://github.com/tontoko/jev-browser","category":"Browser & OS Action","plainSummary":"基于 Playwright 的 Jev 浏览器自动化工具，共用 CLI、MCP 与 TypeScript SDK。","plainSummaryEn":"Playwright browser automation with Jev decisions through a shared CLI, MCP server and TypeScript SDK.","jevDecisionPoint":"Jev 从页面观测中选择操作、匹配表单字段或提取结构化内容，Playwright 执行。","jevDecisionPointEn":"Jev chooses actions, matches form fields or extracts structured content from page observations; Playwright executes.","highlightBenefit":"支持持久会话与页面结果回读；回读不等于已验证数据库持久化。","highlightBenefitEn":"Supports persistent sessions and UI readback, which does not by itself prove database persistence.","tags":["browser-automation","mcp-integrations","multilanguage-sdk"],"stars":6,"forks":1,"openIssues":3,"license":"Apache-2.0","language":"JavaScript","lastCommitAt":"2026-09-21T13:53:48Z","createdAt":"2026-09-17T09:14:40Z","metadataFetchedAt":"2026-09-21T21:55:44.901Z","avatarUrl":"https://avatars.githubusercontent.com/u/31890263?v=4","summarySource":"human-reviewed","claimStatus":"已核对固定版本 README 与调用源码。","claimStatusEn":"Reviewed the fixed-version README and call path.","plainSummaryJa":"Playwright と Jev を共通の CLI・MCP・TypeScript SDK から利用するブラウザー自動化ツール。","plainSummaryKo":"공통 CLI·MCP·TypeScript SDK로 Playwright와 Jev를 사용하는 브라우저 자동화 도구.","jevDecisionPointJa":"ページ観測から Jev が操作・フォーム対応・抽出内容を判断し、Playwright が実行する。","jevDecisionPointKo":"Jev가 페이지 관측으로 동작·양식 필드·추출 내용을 결정하고 Playwright가 실행한다.","highlightBenefitJa":"永続セッションと画面回読に対応。画面確認だけでは DB 永続化の証明にならない。","highlightBenefitKo":"지속 세션과 화면 재확인을 지원하지만 화면 확인만으로 DB 저장을 증명하지는 않는다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/tontoko/jev-browser/blob/05b8257db8d7df77a82274016de84d965a38214f/README.ja.md","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/tontoko/jev-browser/blob/05b8257db8d7df77a82274016de84d965a38214f/README.md","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/tontoko/jev-browser/blob/05b8257db8d7df77a82274016de84d965a38214f/src/decision.ts","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/tontoko/jev-browser/blob/05b8257db8d7df77a82274016de84d965a38214f/src/browser.ts","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/tontoko/jev-browser/blob/05b8257db8d7df77a82274016de84d965a38214f/src/actions.ts","note":"固定版本说明与源码；本站未执行该项目。"}],"pinned":false,"topics":["browser-automation","cli","jev","mcp","playwright","typescript"]},{"id":"theooliveira:pi-jev","name":"pi-jev","author":"TheoOliveira","url":"https://github.com/TheoOliveira/pi-jev","category":"MCP & Integrations","plainSummary":"为 Pi Agent 按任务寻找工具与技能，并提供结构化判断和可选的历史筛选。","plainSummaryEn":"Helps Pi Agents find tools and skills for a task, with structured evaluation and optional history filtering.","jevDecisionPoint":"判断候选工具、技能和历史记录与当前任务的相关性，供本地规则决定加载或保留。","jevDecisionPointEn":"Judges how candidate tools, skills, and history relate to the task so local rules can load or retain them.","highlightBenefit":"让能力按需进入 Agent 工作流。","highlightBenefitEn":"Lets capabilities enter an Agent workflow on demand.","tags":["coding-agents","mcp-integrations","llm-routing-cost"],"stars":31,"forks":3,"openIssues":7,"license":"MIT","language":"TypeScript","lastCommitAt":"2026-09-20T21:08:55Z","createdAt":"2026-09-17T04:00:49Z","metadataFetchedAt":"2026-09-21T21:55:47.951Z","avatarUrl":"https://avatars.githubusercontent.com/u/11945175?v=4","summarySource":"human-reviewed","claimStatus":"这是 Pi 扩展，不能仅因提供工具就称为 MCP server；自动模式需配置。","claimStatusEn":"This is a Pi extension; exposing tools does not by itself make it an MCP server. Automatic modes require configuration.","plainSummaryJa":"Pi Agent に必要なツールとスキルを探し、構造化評価や任意の履歴フィルターを提供する。","plainSummaryKo":"Pi Agent에 필요한 도구와 스킬을 찾고 구조화된 평가와 선택적 기록 필터링을 제공한다.","jevDecisionPointJa":"候補ツール、スキル、履歴とタスクの関連性を判定し、読み込みや保持に使う。","jevDecisionPointKo":"후보 도구, 스킬, 기록의 작업 관련성을 판단해 로딩과 보존에 활용한다.","highlightBenefitJa":"必要な能力を必要なときにワークフローへ加えられる。","highlightBenefitKo":"필요한 기능을 필요할 때 Agent 워크플로에 추가할 수 있다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"confirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/TheoOliveira/pi-jev/blob/0c3ac04bb5e039e7a334e8e79c8ad471a7a1b67c/src/jev.ts","note":"Fixed-commit source reviewed / 已核对固定版本源码"},{"url":"https://github.com/TheoOliveira/pi-jev/blob/0c3ac04bb5e039e7a334e8e79c8ad471a7a1b67c/README.md","note":"Fixed-commit source reviewed / 已核对固定版本源码"}],"pinned":false,"topics":["jev","pi-coding-agent","pi-extension","pi-package","system-one","tool-routing","typesafe"]},{"id":"reachjalil:jev-tree","name":"jev-tree","author":"reachjalil","url":"https://github.com/reachjalil/jev-tree","category":"Classification & Taxonomy","plainSummary":"选项太多，一次问不下？先把目录分成树，让 Jev 逐层选分支，最后落到具体商品、事件类型或工作流。","plainSummaryEn":"A hierarchical selector that asks Jev to traverse branches when a catalog has too many options for one question.","jevDecisionPoint":"每次只在当前层的候选分支里做单选，再沿选中的分支继续查找。","jevDecisionPointEn":"Selects one branch at each level and continues until it reaches a final candidate.","highlightBenefit":"不用直接截掉大目录尾部的候选，还能返回完整选择路径；失败时明确报告不可用。","highlightBenefitEn":"Returns the selected path without silently truncating the tail of a large catalog.","tags":["classification-ranking","typed-decisions"],"stars":3,"forks":0,"openIssues":1,"license":"MIT","language":"TypeScript","lastCommitAt":"2026-09-18T02:12:21Z","createdAt":"2026-09-18T01:03:19Z","metadataFetchedAt":"2026-09-21T21:55:50.935Z","avatarUrl":"https://avatars.githubusercontent.com/u/6254931?v=4","summarySource":"human-reviewed","claimStatus":"分层选择实现可核对；作者评测基于合成数据，本站未重跑真实业务评测。","claimStatusEn":"The hierarchical selection implementation is inspectable; reported evaluations use synthetic data and were not rerun on real workloads here.","plainSummaryJa":"候補が多すぎる目録を階層化し、Jev に枝を順番に選ばせるセレクター。","plainSummaryKo":"후보가 많은 목록을 계층화해 Jev가 가지를 순서대로 고르는 선택기.","jevDecisionPointJa":"各階層で一つの枝を選び、最終候補まで進む。","jevDecisionPointKo":"각 단계에서 가지 하나를 선택하고 최종 후보까지 내려간다.","highlightBenefitJa":"大きな目録の末尾を黙って切り捨てず、選択経路を返す。","highlightBenefitKo":"큰 목록의 뒤쪽을 조용히 잘라내지 않고 선택 경로를 반환한다.","claimStatusJa":"固定バージョンの文書とソースを確認。実行・性能・安全性は当サイトで独立検証していません。","claimStatusKo":"고정 버전 문서와 소스를 확인했습니다. 실행·성능·안전성은 이 사이트에서 독립적으로 검증하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/reachjalil/jev-tree/blob/95bff63bd653fee4dc71f33f9431dce0f81e2ca3/README.md","note":"README.md"},{"url":"https://github.com/reachjalil/jev-tree/blob/95bff63bd653fee4dc71f33f9431dce0f81e2ca3/src/index.ts","note":"src/index.ts"}],"pinned":false,"topics":["classification","jev","taxonomy","typesafe","typescript"]},{"id":"hamakyo:jev-starter","name":"jev-starter","author":"hamakyo","url":"https://github.com/hamakyo/jev-starter","category":"SDK & Decision Frameworks","plainSummary":"在 TypeSafe SDK 上补充决策阈值、备用路径、人工复核和评测模式的 TypeScript 工具集。","plainSummaryEn":"TypeScript patterns for thresholds, fallbacks, human review and evaluation on top of the TypeSafe SDK.","jevDecisionPoint":"Jev 返回结构化判断，应用策略再选择自动处理、fallback 或人工复核。","jevDecisionPointEn":"Jev returns judgments; application policies choose automatic handling, fallback or human review.","highlightBenefit":"把模型调用、执行策略和评测分开组织。","highlightBenefitEn":"Separates model calls, execution policy and evaluation.","tags":["multilanguage-sdk","security-guardrails","evaluation-benchmarks"],"stars":2,"forks":1,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-18T01:43:15Z","createdAt":"2026-09-17T01:48:56Z","metadataFetchedAt":"2026-09-21T21:55:53.933Z","avatarUrl":"https://avatars.githubusercontent.com/u/32177429?v=4","summarySource":"human-reviewed","claimStatus":"已核对固定版本 README 与集成源码。","claimStatusEn":"README and integration source reviewed at a fixed commit; not independently run or benchmarked by this site.","plainSummaryJa":"TypeSafe SDK に、しきい値、代替経路、人による確認、評価のパターンを加える TypeScript ツール集。","plainSummaryKo":"TypeSafe SDK에 임계값, 대체 경로, 사람의 검토와 평가 패턴을 더하는 TypeScript 도구다.","jevDecisionPointJa":"Jev の判断を受け、アプリの規則が自動処理、fallback、人の確認を選ぶ。","jevDecisionPointKo":"Jev 판단을 받아 앱 정책이 자동 처리, fallback 또는 사람의 검토를 고른다.","highlightBenefitJa":"モデル呼び出し、実行方針、評価を分けて構成できる。","highlightBenefitKo":"모델 호출, 실행 정책과 평가를 분리해 구성한다.","claimStatusJa":"固定版の README と実装を確認済み。このサイトでは独立した実行・性能検証をしていません。","claimStatusKo":"고정 버전 README와 구현을 확인했습니다. 이 사이트에서 독립적인 실행·성능 검증은 하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/hamakyo/jev-starter/blob/fb0fae07d1f3052f28af630b942c16b2a7bd8c88/src/providers/jev-provider.ts#L43","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/hamakyo/jev-starter/blob/fb0fae07d1f3052f28af630b942c16b2a7bd8c88/README.md","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/hamakyo/jev-starter/blob/fb0fae07d1f3052f28af630b942c16b2a7bd8c88/README.ja.md","note":"固定版本的实现与说明 / Pinned source and documentation"}],"pinned":false},{"id":"himomohi:aside-jev","name":"aside-jev","author":"himomohi","url":"https://github.com/himomohi/aside-jev","category":"Browser & OS Action","plainSummary":"给 Aside 浏览器 Agent 提供 Jev 判断的 MCP 服务器与 skill。","plainSummaryEn":"An MCP server and skill adding Jev decisions to Aside browser Agents.","jevDecisionPoint":"Agent 列出候选动作，Jev 选择候选 ID，再由 Agent 通过 Aside 执行并复核。","jevDecisionPointEn":"The Agent supplies action candidates; Jev picks an ID, then the Agent executes and verifies it through Aside.","highlightBenefit":"把模型选择限制在应用给出的动作表内，执行结果仍需另行验证。","highlightBenefitEn":"Keeps choices within an application-owned action table while requiring separate outcome verification.","tags":["browser-automation","mcp-integrations"],"stars":6,"forks":3,"openIssues":0,"license":"MIT","language":"Python","lastCommitAt":"2026-09-21T11:31:38Z","createdAt":"2026-09-17T22:07:20Z","metadataFetchedAt":"2026-09-21T21:55:56.972Z","avatarUrl":"https://avatars.githubusercontent.com/u/49967692?v=4","summarySource":"human-reviewed","claimStatus":"已核对固定版本 README 与调用源码。","claimStatusEn":"Reviewed the fixed-version README and call path.","plainSummaryJa":"Aside ブラウザー Agent に Jev 判断を加える MCP サーバーと skill。","plainSummaryKo":"Aside 브라우저 Agent에 Jev 판단을 더하는 MCP 서버와 skill.","jevDecisionPointJa":"Agent が候補を用意し、Jev が ID を選択。Aside で実行した後に結果を確認する。","jevDecisionPointKo":"Agent가 후보를 제시하면 Jev가 ID를 고르고 Aside로 실행한 뒤 검증한다.","highlightBenefitJa":"選択をアプリ側の動作表に限定し、実行結果は別途検証する。","highlightBenefitKo":"선택을 앱이 제공한 동작 목록으로 제한하며 실행 결과는 별도로 확인한다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/himomohi/aside-jev/blob/d2923f7949a9576e82b51d2c5e5514c17d80d60b/README.md","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/himomohi/aside-jev/blob/d2923f7949a9576e82b51d2c5e5514c17d80d60b/src/aside_jev/jev.py","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/himomohi/aside-jev/blob/d2923f7949a9576e82b51d2c5e5514c17d80d60b/src/aside_jev/loop.py","note":"固定版本说明与源码；本站未执行该项目。"}],"pinned":false,"topics":["aside","browser-agent","computer-use","jev","mcp","typesafe"]},{"id":"shivampansuriya:jev-skill-gate","name":"jev-skill-gate","author":"ShivamPansuriya","url":"https://github.com/ShivamPansuriya/jev-skill-gate","category":"Context GC & Filter","plainSummary":"按当前项目给 Claude Code 技能排相关度，减少默认加载的技能说明。","plainSummaryEn":"Ranks Claude Code skills for the current project to reduce the descriptions loaded by default.","jevDecisionPoint":"结合项目技术栈、目录与 README 判断技能相关性，再调整技能说明的可见程度。","jevDecisionPointEn":"Assesses skills against the project stack, directory, and README before adjusting description visibility.","highlightBenefit":"保留需要的说明，同时给其余技能保留手动调用入口。","highlightBenefitEn":"Keeps relevant descriptions while preserving manual access to other skills.","tags":["context-compaction","coding-agents","llm-routing-cost"],"stars":3,"forks":1,"openIssues":1,"license":"MIT","language":"JavaScript","lastCommitAt":"2026-09-17T20:32:51Z","createdAt":"2026-09-17T19:27:46Z","metadataFetchedAt":"2026-09-21T21:55:59.925Z","avatarUrl":"https://avatars.githubusercontent.com/u/121541430?v=4","summarySource":"human-reviewed","claimStatus":"没有 key 时可用本地评分；作者披露 Jev 只覆盖部分评测案例，本站未验证 Token 节省或效果。","claimStatusEn":"Local scoring is available without a key. The author reports only partial Jev evaluation coverage; this site has not verified Token savings or effectiveness.","plainSummaryJa":"現在のプロジェクトに対する Claude Code スキルの関連度を付け、初期表示する説明を絞る。","plainSummaryKo":"현재 프로젝트에 대한 Claude Code 스킬 관련도를 평가해 기본으로 로드할 설명을 줄인다.","jevDecisionPointJa":"技術構成、ディレクトリ、README に照らして関連性を判定し、説明の表示を調整する。","jevDecisionPointKo":"기술 스택, 디렉터리, README에 비춰 관련성을 판단하고 설명 노출 수준을 조정한다.","highlightBenefitJa":"必要な説明を残し、ほかのスキルも手動で呼び出せる。","highlightBenefitKo":"필요한 설명을 유지하면서 나머지 스킬도 수동으로 호출할 수 있다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"confirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/ShivamPansuriya/jev-skill-gate/blob/1ab4c9a7cd117e1c7e129021303416bb7c2f5328/src/jev.mjs","note":"Fixed-commit source reviewed / 已核对固定版本源码"},{"url":"https://github.com/ShivamPansuriya/jev-skill-gate/blob/1ab4c9a7cd117e1c7e129021303416bb7c2f5328/README.md","note":"Fixed-commit source reviewed / 已核对固定版本源码"}],"pinned":false,"topics":["agent-skills","ai-agents","ai-tooling","anthropic","claude","claude-code","claude-code-hooks","cli","context-engineering","context-management","context-window","cost-optimization","developer-tools","jev","llm","llm-tools","nodejs","prompt-engineering","token-optimization","typesafe-ai"]},{"id":"leepokai:jev-guard","name":"jev-guard","author":"leepokai","url":"https://github.com/leepokai/jev-guard","category":"Security & Guardrails","plainSummary":"在编程 Agent 调工具前后加一道检查：操作是否危险、是不是用户要求的、返回内容里有没有诱导 Agent 越界的指令。","plainSummaryEn":"Tool-call checks for coding Agents that assess operation risk, user intent and prompt-injection signals.","jevDecisionPoint":"结合会话对工具风险、用户意图与提示注入迹象做判断，由本地规则决定放行、提醒或拦截。","jevDecisionPointEn":"Jev evaluates tool inputs or outputs in session context; local rules allow, warn or block.","highlightBenefit":"把风险检查接到多种 Agent 的工具流程；确认能力因客户端而异，也不能替代沙箱。","highlightBenefitEn":"Adds review points around Agent tools, with behavior depending on the host client.","tags":["security-guardrails","coding-agents"],"stars":20,"forks":2,"openIssues":3,"license":"MIT","language":"JavaScript","lastCommitAt":"2026-09-17T18:44:09Z","createdAt":"2026-09-17T14:39:28Z","metadataFetchedAt":"2026-09-21T21:56:02.910Z","avatarUrl":"https://avatars.githubusercontent.com/u/109857817?v=4","summarySource":"human-reviewed","claimStatus":"不是沙箱替代品；默认 API 失败时放行，客户端确认能力不同。本站未做安全渗透或端到端复测。","claimStatusEn":"Not a sandbox replacement. API failures allow operations by default and confirmation support varies by client; no penetration or end-to-end testing here.","plainSummaryJa":"コーディング Agent のツール呼び出しで、操作リスク・意図・指示注入の兆候を確認する。","plainSummaryKo":"코딩 Agent의 도구 호출에서 동작 위험·사용자 의도·프롬프트 주입 징후를 검사한다.","jevDecisionPointJa":"Jev が会話文脈で入出力を評価し、ローカル規則が許可・警告・遮断する。","jevDecisionPointKo":"Jev가 대화 맥락에서 입출력을 평가하고 로컬 규칙이 허용·경고·차단한다.","highlightBenefitJa":"ホストごとの制約に合わせて、ツール前後に確認を追加できる。","highlightBenefitKo":"호스트별 제약에 맞춰 도구 실행 전후에 검사 단계를 추가한다.","claimStatusJa":"固定バージョンの文書とソースを確認。実行・性能・安全性は当サイトで独立検証していません。","claimStatusKo":"고정 버전 문서와 소스를 확인했습니다. 실행·성능·안전성은 이 사이트에서 독립적으로 검증하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/leepokai/jev-guard/blob/94996ea80b6b308327ac2077706a29ce6abd3ba0/README.md","note":"README.md"},{"url":"https://github.com/leepokai/jev-guard/blob/94996ea80b6b308327ac2077706a29ce6abd3ba0/src/jev.js","note":"src/jev.js"}],"pinned":false,"topics":["ai-agents","claude-code","codex","cursor","gemini-cli","hooks","jev","opencode","prompt-injection","security"]},{"id":"rashedint32:jev-mcp","name":"jev-mcp","author":"rashedInt32","url":"https://github.com/rashedInt32/jev-mcp","category":"MCP & Integrations","plainSummary":"把 Jev 分类、评分、是非判断和批量提问封装成 MCP 工具，也提供 Claude Code 插件。","plainSummaryEn":"An MCP server and Claude Code plugin for Jev classification, scoring, checks and batched questions.","jevDecisionPoint":"使用 Choice、Score、Noul 调用 TypeSafe 并返回结构化答案。","jevDecisionPointEn":"Calls TypeSafe with Choice, Score and Noul questions and returns structured answers.","highlightBenefit":"让 MCP 客户端直接使用这些判断类型。","highlightBenefitEn":"Makes the typed decision primitives available to MCP clients.","tags":["mcp-integrations","coding-agents","classification-ranking"],"stars":5,"forks":2,"openIssues":1,"license":"MIT","lastCommitAt":"2026-09-20T02:39:58Z","createdAt":"2026-09-17T11:51:54Z","metadataFetchedAt":"2026-09-21T21:56:05.905Z","avatarUrl":"https://avatars.githubusercontent.com/u/8067127?v=4","summarySource":"human-reviewed","claimStatus":"已核对固定版本 README 与集成源码。","claimStatusEn":"README and integration source reviewed at a fixed commit; not independently run or benchmarked by this site.","plainSummaryJa":"分類、採点、真偽判断、一括質問を提供する Jev MCP サーバーと Claude Code プラグイン。","plainSummaryKo":"분류, 점수, 참·거짓 판단과 일괄 질문을 제공하는 Jev MCP 서버 및 Claude Code 플러그인이다.","jevDecisionPointJa":"Choice、Score、Noul で TypeSafe を呼び、構造化された回答を返す。","jevDecisionPointKo":"Choice, Score, Noul로 TypeSafe를 호출해 구조화된 답을 반환한다.","highlightBenefitJa":"MCP クライアントから型付き判断を使える。","highlightBenefitKo":"MCP 클라이언트에서 구조화된 판단 기능을 사용할 수 있다.","claimStatusJa":"固定版の README と実装を確認済み。このサイトでは独立した実行・性能検証をしていません。","claimStatusKo":"고정 버전 README와 구현을 확인했습니다. 이 사이트에서 독립적인 실행·성능 검증은 하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/rashedInt32/jev-mcp/blob/27603c12d06cd7899d93c0fbb8ce9331040b8a8c/src/index.ts#L230","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/rashedInt32/jev-mcp/blob/27603c12d06cd7899d93c0fbb8ce9331040b8a8c/README.md","note":"固定版本的实现与说明 / Pinned source and documentation"}],"pinned":false},{"id":"abovecolin:jevclient","name":"jevclient","author":"AboveColin","url":"https://github.com/AboveColin/jevclient","category":"SDK & Decision Frameworks","plainSummary":"Jev 的异步 Python 客户端，一次请求可提交多个结构化判断问题。","plainSummaryEn":"An asynchronous Python Jev client that batches typed questions in one request.","jevDecisionPoint":"通过 aiohttp 调用 TypeSafe 接口，把分类、分数与概率解析为对象。","jevDecisionPointEn":"Calls TypeSafe through aiohttp and parses choices, scores and probabilities into objects.","highlightBenefit":"便于在现有异步程序中批量提问，不需要解析模型生成的文字。","highlightBenefitEn":"Fits existing async applications without parsing generated prose.","tags":["multilanguage-sdk","typed-decisions"],"stars":2,"forks":0,"openIssues":0,"license":"MIT","language":"Python","lastCommitAt":"2026-09-21T18:39:39Z","createdAt":"2026-09-17T11:00:15Z","metadataFetchedAt":"2026-09-21T21:56:08.975Z","avatarUrl":"https://avatars.githubusercontent.com/u/6614011?v=4","summarySource":"human-reviewed","claimStatus":"已核对固定版本 README 与调用源码。","claimStatusEn":"Reviewed the fixed-version README and call path.","plainSummaryJa":"複数の型付き質問を一度に送れる Jev の非同期 Python クライアント。","plainSummaryKo":"여러 구조화된 질문을 한 번에 보내는 Jev 비동기 Python 클라이언트.","jevDecisionPointJa":"aiohttp で TypeSafe を呼び、選択・スコア・確率をオブジェクトに変換する。","jevDecisionPointKo":"aiohttp로 TypeSafe를 호출해 선택지·점수·확률을 객체로 변환한다.","highlightBenefitJa":"生成文の解析なしで既存の非同期アプリに組み込める。","highlightBenefitKo":"생성된 문장 파싱 없이 기존 비동기 앱에 연결할 수 있다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/AboveColin/jevclient/blob/d568f25717c5b956022f5dc0fcfcb905c7533a05/README.md","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/AboveColin/jevclient/blob/d568f25717c5b956022f5dc0fcfcb905c7533a05/jevclient/client.py","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/AboveColin/jevclient/blob/d568f25717c5b956022f5dc0fcfcb905c7533a05/jevclient/models.py","note":"固定版本说明与源码；本站未执行该项目。"}],"pinned":false,"topics":["async","asyncio","classification","jev","python","typesafe"]},{"id":"molis-ai:jev-workbench","name":"jev-workbench","author":"molis-ai","url":"https://github.com/molis-ai/jev-workbench","category":"MCP & Integrations","plainSummary":"在本地网页定义、试跑与发布 Jev 判断函数，供后端和 Agent 调用固定版本。","plainSummaryEn":"Defines, tests, and publishes Jev decision functions in a local UI so backends and Agents can call fixed versions.","jevDecisionPoint":"把分类、证据判断等条件定义为 Noul、Choice 或 Score 问题后调用 TypeSafe。","jevDecisionPointEn":"Encodes classification or evidence checks as Noul, Choice, or Score questions sent to TypeSafe.","highlightBenefit":"让多个调用方复用同一个有版本的判断函数。","highlightBenefitEn":"Lets multiple callers reuse the same versioned decision function.","tags":["mcp-integrations","typed-decisions","evaluation-benchmarks"],"stars":2,"forks":0,"openIssues":0,"license":"MIT","language":"TypeScript","lastCommitAt":"2026-09-18T04:38:18Z","createdAt":"2026-09-17T18:15:42Z","metadataFetchedAt":"2026-09-21T21:56:12.000Z","avatarUrl":"https://avatars.githubusercontent.com/u/140519386?v=4","summarySource":"human-reviewed","claimStatus":"demo 使用模拟答案，真实调用是另一条 TypeSafe 路径；各 Agent 集成与线上效果未由本站验证。","claimStatusEn":"The demo uses simulated answers; actual TypeSafe calls follow a separate path. This site has not verified each Agent integration or production behavior.","plainSummaryJa":"ローカル画面で Jev 判断関数を定義、試行、公開し、バックエンドや Agent から固定版を呼ぶ。","plainSummaryKo":"로컬 화면에서 Jev 판단 함수를 정의, 시험, 게시하고 백엔드와 Agent에서 고정 버전을 호출한다.","jevDecisionPointJa":"分類や根拠の確認を Noul、Choice、Score として定義し TypeSafe に送る。","jevDecisionPointKo":"분류나 근거 확인을 Noul, Choice, Score 질문으로 정의해 TypeSafe에 보낸다.","highlightBenefitJa":"複数の呼び出し元で同じ版の判断関数を再利用できる。","highlightBenefitKo":"여러 호출자가 동일한 버전의 판단 함수를 재사용할 수 있다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"confirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/molis-ai/jev-workbench/blob/3d7d6f673d629a046f2e3d9b2f277095133a65c7/apps/server/src/provider.ts","note":"Fixed-commit source reviewed / 已核对固定版本源码"},{"url":"https://github.com/molis-ai/jev-workbench/blob/3d7d6f673d629a046f2e3d9b2f277095133a65c7/README.md","note":"Fixed-commit source reviewed / 已核对固定版本源码"}],"pinned":false},{"id":"phureewat29:got-jev","name":"jev-got","author":"phureewat29","url":"https://github.com/phureewat29/jev-got","category":"Creative Tools","plainSummary":"《权力的游戏》文字冒险：大模型写剧情，Jev 判断地点、情绪和危险，再切换配乐与背景。","plainSummaryEn":"A Game of Thrones text-adventure demo where a language model writes the story and Jev labels the scene.","jevDecisionPoint":"对新剧情判断 location、beat、mood、danger、inFiction 五项状态。","jevDecisionPointEn":"Classifies location, narrative beat, mood, danger and whether text remains in fiction.","highlightBenefit":"用固定状态驱动界面和下一回合，不必从生成的故事里硬拆字段。","highlightBenefitEn":"Uses explicit scene state to drive backgrounds, music and the next turn.","tags":["creative-multimedia","games-simulation","classification-ranking"],"stars":2,"forks":0,"openIssues":0,"license":null,"language":"TypeScript","lastCommitAt":"2026-09-19T03:10:11Z","createdAt":"2026-09-17T10:35:36Z","metadataFetchedAt":"2026-09-21T21:56:14.912Z","avatarUrl":"https://avatars.githubusercontent.com/u/2357480?v=4","summarySource":"human-reviewed","claimStatus":"非商业技术 demo；Jev 负责场景分类而非生成故事。游戏。","claimStatusEn":"A noncommercial technical demo: Jev labels scenes rather than generating the story. The game was not run here.","plainSummaryJa":"別の言語モデルが物語を書き、Jev が場面を分類する Game of Thrones の文章ゲーム。","plainSummaryKo":"다른 언어 모델이 이야기를 쓰고 Jev가 장면을 분류하는 Game of Thrones 텍스트 게임.","jevDecisionPointJa":"場所・物語の展開・気分・危険・物語内かどうかを判定する。","jevDecisionPointKo":"장소·서사 흐름·분위기·위험·허구 내 여부를 판단한다.","highlightBenefitJa":"明示的な場面状態で背景・音楽・次のターンを制御する。","highlightBenefitKo":"명시적인 장면 상태로 배경·음악·다음 턴을 제어한다.","claimStatusJa":"固定バージョンの文書とソースを確認。実行・性能・安全性は当サイトで独立検証していません。","claimStatusKo":"고정 버전 문서와 소스를 확인했습니다. 실행·성능·안전성은 이 사이트에서 독립적으로 검증하지 않았습니다.","licenseStatus":"unconfirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/phureewat29/jev-got/blob/232fc0ad0cc6f4d4e095df71ac5f05b0f49033a9/README.md","note":"README.md"},{"url":"https://github.com/phureewat29/jev-got/blob/232fc0ad0cc6f4d4e095df71ac5f05b0f49033a9/src/core/providers/TypeSafe.ts","note":"src/core/providers/TypeSafe.ts"}],"pinned":false},{"id":"joelhooks:pi-fast-jev-compaction","name":"pi-fast-jev-compaction","author":"joelhooks","url":"https://github.com/joelhooks/pi-fast-jev-compaction","category":"Context GC & Filter","plainSummary":"Pi 扩展：清理过时工具历史，保留原文；不足以释放上下文时交给 Pi 原生摘要。","plainSummaryEn":"A Pi extension that prunes stale tool history verbatim and falls back to Pi summary compaction when needed.","jevDecisionPoint":"分别判断调用和结果是否保留，并在送给模型的上下文中执行过滤。","jevDecisionPointEn":"Judges whether calls and outputs should survive in the context sent to the model.","highlightBenefit":"保留原始会话文件，并记录每次删减决定。","highlightBenefitEn":"Retains the original session file and records pruning decisions.","tags":["context-compaction","coding-agents"],"stars":9,"forks":1,"openIssues":1,"license":"MIT","lastCommitAt":"2026-09-18T13:58:40Z","createdAt":"2026-09-18T02:45:06Z","metadataFetchedAt":"2026-09-21T21:56:17.906Z","avatarUrl":"https://avatars.githubusercontent.com/u/86834?v=4","summarySource":"human-reviewed","claimStatus":"已核对固定版本 README 与集成源码。","claimStatusEn":"README and integration source reviewed at a fixed commit; not independently run or benchmarked by this site.","plainSummaryJa":"古いツール履歴を原文のまま整理し、必要なら Pi 標準の要約へ渡す拡張。","plainSummaryKo":"오래된 도구 기록을 정리하고 필요하면 Pi 기본 요약으로 넘기는 확장이다.","jevDecisionPointJa":"モデルに送る履歴で、呼び出しと結果を残す必要があるか判断する。","jevDecisionPointKo":"모델에 보낼 문맥에 호출과 결과를 남길지 판단한다.","highlightBenefitJa":"元のセッションファイルを保持し、削減の判断を記録する。","highlightBenefitKo":"원본 세션 파일을 유지하며 정리 판단을 기록한다.","claimStatusJa":"固定版の README と実装を確認済み。このサイトでは独立した実行・性能検証をしていません。","claimStatusKo":"고정 버전 README와 구현을 확인했습니다. 이 사이트에서 독립적인 실행·성능 검증은 하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/joelhooks/pi-fast-jev-compaction/blob/eb83f533f4fd10a08728b02c062c249afda5a4dc/src/core/request.ts#L8","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/joelhooks/pi-fast-jev-compaction/blob/eb83f533f4fd10a08728b02c062c249afda5a4dc/README.md","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/joelhooks/pi-fast-jev-compaction/blob/eb83f533f4fd10a08728b02c062c249afda5a4dc/src/core/decide.ts","note":"固定版本的实现与说明 / Pinned source and documentation"}],"pinned":false},{"id":"anpicasso:hermes-jev-approvals","name":"hermes-jev-approvals","author":"anpicasso","url":"https://github.com/anpicasso/hermes-jev-approvals","category":"Security & Guardrails","plainSummary":"Hermes 的实验性命令审批插件，只替换 auxiliary.approval 判断任务。","plainSummaryEn":"An experimental Hermes command-approval provider for the auxiliary.approval task only.","jevDecisionPoint":"Jev 返回 APPROVE、DENY 或 ESCALATE，由本地阈值和策略决定最终处理。","jevDecisionPointEn":"Jev returns APPROVE, DENY or ESCALATE, with local thresholds and policy determining handling.","highlightBenefit":"展示审批接入方式；作者明确标为概念验证，不能当作生产安全保证。","highlightBenefitEn":"Demonstrates an approval integration; the author labels it a proof of concept, not production assurance.","tags":["security-guardrails","mcp-integrations"],"stars":12,"forks":2,"openIssues":1,"license":"MIT","language":"Python","lastCommitAt":"2026-09-20T19:55:39Z","createdAt":"2026-09-17T20:33:17Z","metadataFetchedAt":"2026-09-21T21:56:20.907Z","avatarUrl":"https://avatars.githubusercontent.com/u/24831723?v=4","summarySource":"human-reviewed","claimStatus":"作者标注概念验证，仅离线基准。","claimStatusEn":"The author labels this a proof of concept with offline benchmarks.","plainSummaryJa":"Hermes の auxiliary.approval だけを置き換える実験的なコマンド承認プラグイン。","plainSummaryKo":"Hermes의 auxiliary.approval만 대체하는 실험적 명령 승인 플러그인.","jevDecisionPointJa":"Jev の APPROVE・DENY・ESCALATE をローカル閾値と方針で処理する。","jevDecisionPointKo":"Jev의 APPROVE·DENY·ESCALATE를 로컬 임곗값과 정책으로 처리한다.","highlightBenefitJa":"承認への接続例。作者が概念実証と明記しており、本番の安全性を保証しない。","highlightBenefitKo":"승인 연동 예제이며 저자가 개념 증명으로 명시해 운영 보안을 보장하지 않는다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/anpicasso/hermes-jev-approvals/blob/c41d81de23a79d8d69397cef81c4de99b3f987c5/README.md","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/anpicasso/hermes-jev-approvals/blob/c41d81de23a79d8d69397cef81c4de99b3f987c5/plugin/__init__.py","note":"固定版本说明与源码；本站未执行该项目。"}],"pinned":false},{"id":"sherajdev:jev-research","name":"jev-research","author":"sherajdev","url":"https://github.com/sherajdev/jev-research","category":"Routing & Cost Optimization","plainSummary":"一份 Jev 与 Herdr 协作指南，附有把任务分给不同 Agent 的路由原型。","plainSummaryEn":"A Jev and Herdr integration guide with a prototype for routing tasks to different Agents.","jevDecisionPoint":"根据任务与仓库状态选择执行者，并给出风险和派发准备情况的判断。","jevDecisionPointEn":"Uses task and repository state to choose an executor and assess risk and dispatch readiness.","highlightBenefit":"提供可以阅读和修改的多 Agent 派工示例。","highlightBenefitEn":"Provides a readable, editable example of multi-Agent task routing.","tags":["llm-routing-cost","typed-decisions"],"stars":1,"forks":0,"openIssues":0,"license":"MIT","language":"TypeScript","lastCommitAt":"2026-09-18T02:23:09Z","createdAt":"2026-09-18T02:15:07Z","metadataFetchedAt":"2026-09-21T21:56:23.948Z","avatarUrl":"https://avatars.githubusercontent.com/u/228158241?v=4","summarySource":"human-reviewed","claimStatus":"主要是文档与 dry-run 路由原型，不能当作已验证运行的完整调度平台。","claimStatusEn":"Primarily documentation and a dry-run routing prototype, not a verified operational orchestration platform.","plainSummaryJa":"Jev と Herdr の連携ガイドと、タスクを各 Agent に振り分ける試作ルーター。","plainSummaryKo":"Jev와 Herdr 연동 가이드와 작업을 여러 Agent에 배분하는 라우터 프로토타입.","jevDecisionPointJa":"タスクとリポジトリ状態から実行先、リスク、派遣の準備状況を判断する。","jevDecisionPointKo":"작업과 저장소 상태로 실행자, 위험, 배분 준비 여부를 판단한다.","highlightBenefitJa":"複数 Agent への仕事の割り当てを学べる編集可能な例。","highlightBenefitKo":"여러 Agent에 작업을 배분하는 읽고 수정 가능한 예제를 제공한다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"confirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/sherajdev/jev-research/blob/23da0defeebe5dae232ba1aaa7ee3c0c3842cb8e/jev-router.ts","note":"Fixed-commit source reviewed / 已核对固定版本源码"},{"url":"https://github.com/sherajdev/jev-research/blob/23da0defeebe5dae232ba1aaa7ee3c0c3842cb8e/README.md","note":"Fixed-commit source reviewed / 已核对固定版本源码"}],"pinned":false},{"id":"vicente-md:jev-resilience","name":"jev-resilience","author":"Vicente-MD","url":"https://github.com/Vicente-MD/jev-resilience","category":"SDK & Integrations","plainSummary":"给 Spring WebFlux 检查“HTTP 200 但正文其实报错”的响应。","plainSummaryEn":"A Spring WebFlux integration that detects error messages hidden in HTTP 200 response bodies.","jevDecisionPoint":"Jev 判断响应正文是否是隐藏错误或维护通知，超过阈值就抛出业务异常。","jevDecisionPointEn":"Jev checks for hidden errors or maintenance notices; a local threshold raises a business exception.","highlightBenefit":"让现有错误处理能看到状态码掩盖的失败。","highlightBenefitEn":"Lets existing error-handling code see failures concealed by the HTTP status.","tags":["multilanguage-sdk","evaluation-benchmarks","typed-decisions"],"stars":2,"forks":0,"openIssues":0,"license":null,"language":"Java","lastCommitAt":"2026-09-17T23:52:03Z","createdAt":"2026-09-17T19:26:12Z","metadataFetchedAt":"2026-09-21T21:56:26.929Z","avatarUrl":"https://avatars.githubusercontent.com/u/103980154?v=4","summarySource":"human-reviewed","claimStatus":"确有 Jev API 接入；API 失败时放行，非阻塞不等于没有网络延迟。","claimStatusEn":"Jev API integration is present. API errors fail open, and nonblocking code still incurs network latency.","plainSummaryJa":"HTTP 200 の本文に隠れたエラーを検出する Spring WebFlux 向け連携。","plainSummaryKo":"HTTP 200 응답 본문에 숨은 오류를 감지하는 Spring WebFlux 연동.","jevDecisionPointJa":"Jev がエラーやメンテナンス通知を判定し、閾値に従って例外を発生させる。","jevDecisionPointKo":"Jev가 오류나 점검 공지를 판단하고 임계값에 따라 예외를 발생시킨다.","highlightBenefitJa":"HTTP ステータスに現れない失敗を既存のエラー処理へ渡せる。","highlightBenefitKo":"HTTP 상태에 드러나지 않는 실패를 기존 오류 처리에 전달한다.","claimStatusJa":"固定バージョンの文書とソースを確認。実行・性能・安全性は当サイトで独立検証していません。","claimStatusKo":"고정 버전 문서와 소스를 확인했습니다. 실행·성능·안전성은 이 사이트에서 독립적으로 검증하지 않았습니다.","licenseStatus":"unconfirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/Vicente-MD/jev-resilience/blob/c490e0dc7830758f84bd9d5acb806655e327113e/README.md","note":"README.md"},{"url":"https://github.com/Vicente-MD/jev-resilience/blob/c490e0dc7830758f84bd9d5acb806655e327113e/src/main/java/ai/jev/resilience/client/dto/JevRequest.java","note":"src/main/java/ai/jev/resilience/client/dto/JevRequest.java"},{"url":"https://github.com/Vicente-MD/jev-resilience/blob/c490e0dc7830758f84bd9d5acb806655e327113e/src/main/java/ai/jev/resilience/client/JevEvaluationService.java","note":"src/main/java/ai/jev/resilience/client/JevEvaluationService.java"}],"pinned":false},{"id":"stumble:jev-go","name":"jev-go","author":"Stumble","url":"https://github.com/Stumble/jev-go","category":"SDK & Decision Frameworks","plainSummary":"社区 Go SDK 与命令行，支持 TypeSafe 直连和 Vercel AI Gateway。","plainSummaryEn":"A community Go SDK and CLI supporting TypeSafe directly and Vercel AI Gateway.","jevDecisionPoint":"把应用状态和问题发送给 Jev，解析 Choice、Score 或 Noul 答案。","jevDecisionPointEn":"Sends application state and questions to Jev and parses Choice, Score or Noul answers.","highlightBenefit":"封装请求和响应解析，便于在应用中接入结构化判断。","highlightBenefitEn":"Packages requests and response parsing for typed decisions in applications.","tags":["multilanguage-sdk","cli-git-gates"],"stars":4,"forks":0,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-17T22:56:42Z","createdAt":"2026-09-17T20:55:27Z","metadataFetchedAt":"2026-09-21T21:56:29.932Z","avatarUrl":"https://avatars.githubusercontent.com/u/3461891?v=4","summarySource":"human-reviewed","claimStatus":"已核对固定版本 README 与集成源码。","claimStatusEn":"README and integration source reviewed at a fixed commit; not independently run or benchmarked by this site.","plainSummaryJa":"TypeSafe 直結と Vercel AI Gateway に対応するコミュニティ製 Go SDK と CLI。","plainSummaryKo":"TypeSafe 직접 연결과 Vercel AI Gateway를 지원하는 커뮤니티 Go SDK 및 CLI다.","jevDecisionPointJa":"状態と質問を Jev に送り、Choice、Score、Noul の回答を解析する。","jevDecisionPointKo":"상태와 질문을 Jev에 보내 Choice, Score, Noul 답변을 처리한다.","highlightBenefitJa":"リクエストと応答解析をまとめ、アプリに型付き判断を組み込める。","highlightBenefitKo":"요청과 응답 처리를 묶어 앱에 구조화된 판단을 연결한다.","claimStatusJa":"固定版の README と実装を確認済み。このサイトでは独立した実行・性能検証をしていません。","claimStatusKo":"고정 버전 README와 구현을 확인했습니다. 이 사이트에서 독립적인 실행·성능 검증은 하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/Stumble/jev-go/blob/a475dc925ba68602be93f4478e1381cf5ec27ee4/client.go#L23","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/Stumble/jev-go/blob/a475dc925ba68602be93f4478e1381cf5ec27ee4/README.md","note":"固定版本的实现与说明 / Pinned source and documentation"}],"pinned":false},{"id":"tokentrim:jev-agent-failure-benchmark","name":"jev-agent-failure-benchmark","author":"TokenTrim","url":"https://github.com/TokenTrim/jev-agent-failure-benchmark","category":"Evaluation & Observability","plainSummary":"用 Jev 分析多 Agent 失败记录的评测项目，预测责任 Agent、关键步骤和错误类型。","plainSummaryEn":"A benchmark using Jev to attribute multi-Agent failures to an Agent, step and error type.","jevDecisionPoint":"把记录里的 Agent、步骤和错误分类作为选项，提交三组 choice 问题。","jevDecisionPointEn":"Builds candidate sets from traces and submits three choice questions.","highlightBenefit":"提供评测脚本和作者结果；部分基线生成答案而 Jev 选候选，比较条件不同。","highlightBenefitEn":"Provides evaluation scripts and author results; some baselines generate answers while Jev selects candidates.","tags":["evaluation-benchmarks","classification-ranking"],"stars":2,"forks":0,"openIssues":1,"license":"Apache-2.0","language":"Python","lastCommitAt":"2026-09-17T22:18:39Z","createdAt":"2026-09-17T14:03:46Z","metadataFetchedAt":"2026-09-21T21:56:32.935Z","avatarUrl":"https://avatars.githubusercontent.com/u/269753402?v=4","summarySource":"human-reviewed","claimStatus":"作者的 benchmark 中，Jev 部分任务采用候选选择，论文基线采用自由生成，不能把全部指标当作严格同条件比较。","claimStatusEn":"Some Jev benchmark axes use constrained choices while paper baselines generate freely; not every metric is a like-for-like comparison.","plainSummaryJa":"複数 Agent の失敗記録から責任 Agent・重要ステップ・エラー種別を予測する評価プロジェクト。","plainSummaryKo":"다중 Agent 실패 기록에서 책임 Agent·핵심 단계·오류 유형을 예측하는 평가 프로젝트.","jevDecisionPointJa":"記録から候補集合を作り、Jev に三つの choice 質問を送る。","jevDecisionPointKo":"기록에서 후보 집합을 만들고 Jev에 세 가지 choice 질문을 보낸다.","highlightBenefitJa":"評価スクリプトと作者結果を公開。一部ベースラインの自由生成と候補選択では条件が異なる。","highlightBenefitKo":"평가 코드와 저자 결과를 제공하나 일부 기준 모델의 자유 생성과 후보 선택은 조건이 다르다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/TokenTrim/jev-agent-failure-benchmark/blob/4d46af795a4a4409940a65857da73e45abaea2db/README.md","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/TokenTrim/jev-agent-failure-benchmark/blob/4d46af795a4a4409940a65857da73e45abaea2db/src/jevbench/backends/jev.py","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/TokenTrim/jev-agent-failure-benchmark/blob/4d46af795a4a4409940a65857da73e45abaea2db/RESULTS.md","note":"固定版本说明与源码；本站未执行该项目。"}],"pinned":false},{"id":"paulsmith:computer-use-jev","name":"computer-use-jev","author":"paulsmith","url":"https://github.com/paulsmith/computer-use-jev","category":"Browser & OS Action","plainSummary":"用 Go 控制 macOS 应用，让 Jev 从可访问性树里选择控件和操作。","plainSummaryEn":"A Go-based macOS controller that asks Jev to choose controls and actions from the accessibility tree.","jevDecisionPoint":"根据窗口状态选择动作、目标、是否输入文字以及任务是否结束。","jevDecisionPointEn":"Uses window state to select an action, target, text-input need, and completion status.","highlightBenefit":"候选控件来自界面快照，可以回看选择过程。","highlightBenefitEn":"Grounds candidate controls in UI snapshots so choices can be inspected.","tags":["desktop-os","typed-decisions"],"stars":3,"forks":0,"openIssues":0,"license":"MIT","language":"Go","lastCommitAt":"2026-09-17T19:34:18Z","createdAt":"2026-09-16T22:26:38Z","metadataFetchedAt":"2026-09-21T21:56:35.913Z","avatarUrl":"https://avatars.githubusercontent.com/u/1210?v=4","summarySource":"human-reviewed","claimStatus":"需要 macOS 权限与 TypeSafe key；离线假 worker/HTTP 测试不等于真实桌面验证。","claimStatusEn":"Requires macOS permissions and a TypeSafe key. Offline fake-worker and HTTP tests are not real desktop validation.","plainSummaryJa":"macOS のアクセシビリティツリーから Jev が対象と操作を選ぶ Go 製のコントローラー。","plainSummaryKo":"macOS 접근성 트리에서 Jev가 대상과 동작을 선택하는 Go 컨트롤러.","jevDecisionPointJa":"ウィンドウ状態から操作、対象、文字入力の要否、完了状態を選ぶ。","jevDecisionPointKo":"창 상태로 동작, 대상, 텍스트 입력 필요 여부, 완료 상태를 선택한다.","highlightBenefitJa":"画面スナップショットに基づく候補と選択過程を確認できる。","highlightBenefitKo":"화면 스냅샷에 기반한 후보와 선택 과정을 확인할 수 있다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"confirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/paulsmith/computer-use-jev/blob/ff0ad8ba8e25755d37fb8d93718144c4568a596b/typesafe/client.go","note":"Fixed-commit source reviewed / 已核对固定版本源码"},{"url":"https://github.com/paulsmith/computer-use-jev/blob/ff0ad8ba8e25755d37fb8d93718144c4568a596b/README.md","note":"Fixed-commit source reviewed / 已核对固定版本源码"}],"pinned":false},{"id":"4anti:jev-broadcast-lab","name":"jev-broadcast-lab","author":"4anti","url":"https://github.com/4anti/jev-broadcast-lab","category":"High-Frequency & Simulation","plainSummary":"一个以国际象棋为主的 Jev 实验台，也能试工单分类、文档匹配和审核。","plainSummaryEn":"A Jev experiment workbench centered on chess, with additional classification and matching exercises.","jevDecisionPoint":"从 chess.js 算出的合法走法中选择一步；Stockfish 分数只给操作者看。","jevDecisionPointEn":"Chooses from locally generated legal chess moves; Stockfish scores are shown to the operator rather than Jev.","highlightBenefit":"可以同时观察模型选步和本地棋力评估，检查两者的差别。","highlightBenefitEn":"Lets users compare model move choices with a separate local chess evaluation.","tags":["games-simulation","evaluation-benchmarks","classification-ranking"],"stars":2,"forks":0,"openIssues":0,"license":null,"language":"JavaScript","lastCommitAt":"2026-09-20T10:56:45Z","createdAt":"2026-09-17T16:55:16Z","metadataFetchedAt":"2026-09-21T21:56:38.994Z","avatarUrl":"https://avatars.githubusercontent.com/u/66317321?v=4","summarySource":"human-reviewed","claimStatus":"实验项目；走法先经本地启发式筛选，显示的 Elo 不是认证棋力。","claimStatusEn":"An experiment with local heuristic candidate filtering; displayed Elo is not a certified playing strength. Not run here.","plainSummaryJa":"チェスを中心に、分類や照合も試せる Jev 実験ワークベンチ。","plainSummaryKo":"체스를 중심으로 분류와 매칭도 시험하는 Jev 실험 도구.","jevDecisionPointJa":"ローカルで生成した合法手から選ぶ。Stockfish の評価は操作者に表示する。","jevDecisionPointKo":"로컬에서 생성한 합법 수 중 하나를 선택하며 Stockfish 평가는 사용자에게 표시한다.","highlightBenefitJa":"モデルの選択と別のチェス評価を並べて比較できる。","highlightBenefitKo":"모델이 고른 수와 별도의 체스 평가를 비교할 수 있다.","claimStatusJa":"固定バージョンの文書とソースを確認。実行・性能・安全性は当サイトで独立検証していません。","claimStatusKo":"고정 버전 문서와 소스를 확인했습니다. 실행·성능·안전성은 이 사이트에서 독립적으로 검증하지 않았습니다.","licenseStatus":"unconfirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/4anti/jev-broadcast-lab/blob/90ebea44467489e03c1fd2ab0416b54e32c25c13/README.md","note":"README.md"},{"url":"https://github.com/4anti/jev-broadcast-lab/blob/90ebea44467489e03c1fd2ab0416b54e32c25c13/web/shared/jev-client.js","note":"web/shared/jev-client.js"}],"pinned":false},{"id":"fgn:jevgo","name":"jevgo","author":"fgn","url":"https://github.com/fgn/jevgo","category":"SDK & Decision Frameworks","plainSummary":"社区 Go 客户端，核心只依赖标准库，另有可选 Langfuse 追踪模块。","plainSummaryEn":"A community Go client with a standard-library core and optional Langfuse tracing.","jevDecisionPoint":"把应用状态和问题发送给 Jev，解析 Choice、Score 或 Noul 答案。","jevDecisionPointEn":"Sends application state and questions to Jev and parses Choice, Score or Noul answers.","highlightBenefit":"封装请求和响应解析，便于在应用中接入结构化判断。","highlightBenefitEn":"Packages requests and response parsing for typed decisions in applications.","tags":["multilanguage-sdk","evaluation-benchmarks","typed-decisions"],"stars":2,"forks":0,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-17T15:56:52Z","createdAt":"2026-09-17T14:03:36Z","metadataFetchedAt":"2026-09-21T21:56:41.961Z","avatarUrl":"https://avatars.githubusercontent.com/u/8796563?v=4","summarySource":"human-reviewed","claimStatus":"已核对固定版本 README 与集成源码。","claimStatusEn":"README and integration source reviewed at a fixed commit; not independently run or benchmarked by this site.","plainSummaryJa":"標準ライブラリのみのコアと、任意の Langfuse 追跡を備える Go クライアント。","plainSummaryKo":"표준 라이브러리 기반 코어와 선택형 Langfuse 추적을 제공하는 Go 클라이언트다.","jevDecisionPointJa":"状態と質問を Jev に送り、Choice、Score、Noul の回答を解析する。","jevDecisionPointKo":"상태와 질문을 Jev에 보내 Choice, Score, Noul 답변을 처리한다.","highlightBenefitJa":"リクエストと応答解析をまとめ、アプリに型付き判断を組み込める。","highlightBenefitKo":"요청과 응답 처리를 묶어 앱에 구조화된 판단을 연결한다.","claimStatusJa":"固定版の README と実装を確認済み。このサイトでは独立した実行・性能検証をしていません。","claimStatusKo":"고정 버전 README와 구현을 확인했습니다. 이 사이트에서 독립적인 실행·성능 검증은 하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/fgn/jevgo/blob/ff7a543dda89cbd93e44d6d6535b1a7a62ffb882/client.go#L18","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/fgn/jevgo/blob/ff7a543dda89cbd93e44d6d6535b1a7a62ffb882/README.md","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/fgn/jevgo/blob/ff7a543dda89cbd93e44d6d6535b1a7a62ffb882/contrib/langfuse/README.md","note":"固定版本的实现与说明 / Pinned source and documentation"}],"pinned":false},{"id":"lbotinelly:jev-little-airways","name":"jev-little-airways","author":"lbotinelly","url":"https://github.com/lbotinelly/jev-little-airways","category":"High-Frequency & Simulation","plainSummary":"小岛机场模拟器，让 Jev 判断飞机航路、避让、紧急广播与降落顺序。","plainSummaryEn":"An island-airport simulator using Jev for routes, yielding, emergency broadcasts and landing order.","jevDecisionPoint":"把飞机和周边交通状态转为问题，Jev 的回答驱动模拟器动作。","jevDecisionPointEn":"Turns aircraft and nearby traffic state into questions whose answers drive simulated actions.","highlightBenefit":"可查看请求、回答与模拟状态；仓库也提供 mock 路径。","highlightBenefitEn":"Shows requests, answers and simulation state, with a separate mock path available.","tags":["games-simulation","typed-decisions"],"stars":5,"forks":0,"openIssues":0,"license":"MIT","language":"HTML","lastCommitAt":"2026-09-17T14:19:04Z","createdAt":"2026-09-17T13:17:49Z","metadataFetchedAt":"2026-09-21T21:56:44.937Z","avatarUrl":"https://avatars.githubusercontent.com/u/6088476?v=4","summarySource":"human-reviewed","claimStatus":"已核对固定版本 README 与调用源码。","claimStatusEn":"Reviewed the fixed-version README and call path.","plainSummaryJa":"Jev が航路・譲り合い・緊急放送・着陸順を判断する島の空港シミュレーター。","plainSummaryKo":"Jev가 항로·양보·비상 방송·착륙 순서를 판단하는 섬 공항 시뮬레이터.","jevDecisionPointJa":"機体と周辺交通の状態を質問にし、回答を模擬動作に反映する。","jevDecisionPointKo":"항공기와 주변 교통 상태를 질문으로 만들고 응답을 시뮬레이션 동작에 반영한다.","highlightBenefitJa":"要求・回答・状態を確認でき、別途 mock 経路も備える。","highlightBenefitKo":"요청·응답·상태를 확인할 수 있으며 별도 mock 경로도 제공한다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/lbotinelly/jev-little-airways/blob/d6ec286c7821f41228bce9b33381546e4fca9bff/README.md","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/lbotinelly/jev-little-airways/blob/d6ec286c7821f41228bce9b33381546e4fca9bff/demo/server.mjs","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/lbotinelly/jev-little-airways/blob/d6ec286c7821f41228bce9b33381546e4fca9bff/demo/js/jev-provider.mjs","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/lbotinelly/jev-little-airways/blob/d6ec286c7821f41228bce9b33381546e4fca9bff/demo/js/browser.mjs","note":"固定版本说明与源码；本站未执行该项目。"}],"pinned":false},{"id":"little-planet-labs:jev-playground","name":"jev-playground","author":"Little-Planet-Labs","url":"https://github.com/Little-Planet-Labs/jev-playground","category":"Decision Tools","plainSummary":"在网页里输入状态与选择题或评分题，查看 Jev 的回答和概率分布。","plainSummaryEn":"A web playground for entering state and decision questions, then inspecting Jev answers and probability distributions.","jevDecisionPoint":"把多道 Noul、Choice、Score 问题放在同一次请求里。","jevDecisionPointEn":"Submits multiple Noul, Choice, and Score questions in one request.","highlightBenefit":"不用先写业务代码就能尝试问题与候选答案。","highlightBenefitEn":"Lets users try questions and answer options before writing application code.","tags":["typed-decisions","evaluation-benchmarks"],"stars":1,"forks":0,"openIssues":0,"license":null,"language":"TypeScript","lastCommitAt":"2026-09-17T13:56:19Z","createdAt":"2026-09-17T13:30:44Z","metadataFetchedAt":"2026-09-21T21:56:47.933Z","avatarUrl":"https://avatars.githubusercontent.com/u/325727261?v=4","summarySource":"human-reviewed","claimStatus":"需要真实 TypeSafe key 才调用模型。","claimStatusEn":"A real TypeSafe key is required for model calls. This site has not run inference or verified throughput or accuracy.","plainSummaryJa":"状態や選択・採点の質問を入力し、Jev の回答と確率分布を見る Web 実験画面。","plainSummaryKo":"상태와 선택·평가 질문을 입력하고 Jev 답변과 확률 분포를 살펴보는 웹 실험 도구.","jevDecisionPointJa":"複数の Noul、Choice、Score を一つのリクエストにまとめる。","jevDecisionPointKo":"여러 Noul, Choice, Score 질문을 하나의 요청에 담는다.","highlightBenefitJa":"アプリのコードを書く前に質問と選択肢を試せる。","highlightBenefitKo":"애플리케이션 코드를 작성하기 전에 질문과 선택지를 시험할 수 있다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"unconfirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/Little-Planet-Labs/jev-playground/blob/418b723779569b1cb1f21a4f7f34c546d44c7e34/src/app/api/evaluate/route.ts","note":"Fixed-commit source reviewed / 已核对固定版本源码"},{"url":"https://github.com/Little-Planet-Labs/jev-playground/blob/418b723779569b1cb1f21a4f7f34c546d44c7e34/README.md","note":"Fixed-commit source reviewed / 已核对固定版本源码"}],"pinned":false},{"id":"butochnikov:laravel-typesafe-jev","name":"laravel-typesafe-jev","author":"Butochnikov","url":"https://github.com/Butochnikov/laravel-typesafe-jev","category":"SDK & Integrations","plainSummary":"把 Jev 接入 Laravel，提供配置、依赖注入、Facade 和可记录请求的测试替身。","plainSummaryEn":"A Laravel adapter for Jev with configuration, dependency injection, a Facade and a request-recording test fake.","jevDecisionPoint":"复用社区 PHP SDK 发出三类判断请求，保留类型、异步 Promise 和异常。","jevDecisionPointEn":"Wraps a community PHP SDK for typed judgments and preserves promises and exceptions.","highlightBenefit":"Laravel 服务与队列任务可以沿用自己的配置和测试方式。","highlightBenefitEn":"Fits Jev calls into Laravel services, queues and existing tests.","tags":["multilanguage-sdk","typed-decisions"],"stars":2,"forks":1,"openIssues":0,"license":"MIT","language":"PHP","lastCommitAt":"2026-09-17T13:29:40Z","createdAt":"2026-09-17T12:00:17Z","metadataFetchedAt":"2026-09-21T21:56:50.933Z","avatarUrl":"https://avatars.githubusercontent.com/u/4212297?v=4","summarySource":"human-reviewed","claimStatus":"非官方框架适配；fake 不联网，模拟测试不代表真实 API 通过。","claimStatusEn":"An unofficial framework adapter. Its fake makes no network calls, so mock tests do not validate the live API.","plainSummaryJa":"設定・依存性注入・Facade・リクエスト記録用 fake を備えた Laravel 向け Jev アダプター。","plainSummaryKo":"설정·의존성 주입·Facade·요청 기록용 fake를 갖춘 Laravel용 Jev 어댑터.","jevDecisionPointJa":"コミュニティ製 PHP SDK を包み、型付き判断・Promise・例外を扱う。","jevDecisionPointKo":"커뮤니티 PHP SDK를 감싸 타입이 있는 판단·Promise·예외를 처리한다.","highlightBenefitJa":"Laravel のサービス・キュー・テストへ Jev を組み込める。","highlightBenefitKo":"Laravel 서비스·큐·기존 테스트에 Jev를 연결한다.","claimStatusJa":"固定バージョンの文書とソースを確認。実行・性能・安全性は当サイトで独立検証していません。","claimStatusKo":"고정 버전 문서와 소스를 확인했습니다. 실행·성능·안전성은 이 사이트에서 독립적으로 검증하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/Butochnikov/laravel-typesafe-jev/blob/4d70f5e1b059aac2f63105e72626802247a51f45/README.md","note":"README.md"},{"url":"https://github.com/Butochnikov/laravel-typesafe-jev/blob/4d70f5e1b059aac2f63105e72626802247a51f45/src/JevServiceProvider.php","note":"src/JevServiceProvider.php"},{"url":"https://github.com/Butochnikov/laravel-typesafe-jev/blob/4d70f5e1b059aac2f63105e72626802247a51f45/src/Jev.php","note":"src/Jev.php"}],"pinned":false},{"id":"guillemus:jev-go","name":"jev-go","author":"guillemus","url":"https://github.com/guillemus/jev-go","category":"SDK & Decision Frameworks","plainSummary":"接口精简的非官方 Go SDK，可调用 Jev 并列出可用模型。","plainSummaryEn":"A small unofficial Go SDK for Jev calls and model listing.","jevDecisionPoint":"把应用状态和问题发送给 Jev，解析 Choice、Score 或 Noul 答案。","jevDecisionPointEn":"Sends application state and questions to Jev and parses Choice, Score or Noul answers.","highlightBenefit":"封装请求和响应解析，便于在应用中接入结构化判断。","highlightBenefitEn":"Packages requests and response parsing for typed decisions in applications.","tags":["multilanguage-sdk","typed-decisions"],"stars":1,"forks":0,"openIssues":0,"license":null,"lastCommitAt":"2026-09-17T10:03:41Z","createdAt":"2026-09-17T10:05:24Z","metadataFetchedAt":"2026-09-21T21:56:54.049Z","avatarUrl":"https://avatars.githubusercontent.com/u/202654314?v=4","summarySource":"human-reviewed","claimStatus":"已核对固定版本 README 与集成源码。","claimStatusEn":"README and integration source reviewed at a fixed commit; not independently run or benchmarked by this site.","plainSummaryJa":"Jev 呼び出しとモデル一覧取得に対応する小さな非公式 Go SDK。","plainSummaryKo":"Jev 호출과 모델 목록 조회를 지원하는 간결한 비공식 Go SDK다.","jevDecisionPointJa":"状態と質問を Jev に送り、Choice、Score、Noul の回答を解析する。","jevDecisionPointKo":"상태와 질문을 Jev에 보내 Choice, Score, Noul 답변을 처리한다.","highlightBenefitJa":"リクエストと応答解析をまとめ、アプリに型付き判断を組み込める。","highlightBenefitKo":"요청과 응답 처리를 묶어 앱에 구조화된 판단을 연결한다.","claimStatusJa":"固定版の README と実装を確認済み。このサイトでは独立した実行・性能検証をしていません。","claimStatusKo":"고정 버전 README와 구현을 확인했습니다. 이 사이트에서 독립적인 실행·성능 검증은 하지 않았습니다.","licenseStatus":"unconfirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/guillemus/jev-go/blob/06df95fe081e8b05bf4c0db8292ceefd995fee84/jev.go#L18","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/guillemus/jev-go/blob/06df95fe081e8b05bf4c0db8292ceefd995fee84/README.md","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/guillemus/jev-go/blob/06df95fe081e8b05bf4c0db8292ceefd995fee84/examples/README.md","note":"固定版本的实现与说明 / Pinned source and documentation"}],"pinned":false},{"id":"omni-:ask-jev","name":"ask-jev","author":"omni-","url":"https://github.com/omni-/ask-jev","category":"Evaluation & Observability","plainSummary":"Windows PowerShell 工具，在 Codex 中用 :jev 审视会话里已记录的执行证据。","plainSummaryEn":"A Windows PowerShell tool for auditing recorded Codex execution evidence with :jev.","jevDecisionPoint":"把选定记录交给 Jev，判断执行结论和证据是否充分，输出概率。","jevDecisionPointEn":"Sends selected records to Jev for judgments about execution claims and evidence sufficiency.","highlightBenefit":"显式触发才读取并发送记录；判断供参考，不替代实际测试。","highlightBenefitEn":"Only reads and sends records on explicit invocation; judgments do not replace real tests.","tags":["coding-agents","cli-git-gates","evaluation-benchmarks"],"stars":1,"forks":0,"openIssues":0,"license":"MIT","language":"PowerShell","lastCommitAt":"2026-09-16T23:43:44Z","createdAt":"2026-09-16T23:21:42Z","metadataFetchedAt":"2026-09-21T21:56:56.928Z","avatarUrl":"https://avatars.githubusercontent.com/u/4252190?v=4","summarySource":"human-reviewed","claimStatus":"已核对固定版本 README 与调用源码。","claimStatusEn":"Reviewed the fixed-version README and call path.","plainSummaryJa":"Codex で :jev を使い、記録済みの実行証拠を点検する Windows PowerShell ツール。","plainSummaryKo":"Codex에서 :jev로 기록된 실행 증거를 점검하는 Windows PowerShell 도구.","jevDecisionPointJa":"選択した記録を Jev に送り、結論と証拠の十分さを確率で判断する。","jevDecisionPointKo":"선택한 기록을 Jev에 보내 결론과 증거 충분성을 확률로 판단한다.","highlightBenefitJa":"明示的に呼んだ時だけ記録を読み送信する。実テストの代替ではない。","highlightBenefitKo":"명시적으로 호출할 때만 기록을 읽고 전송하며 실제 테스트를 대체하지 않는다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/omni-/ask-jev/blob/74199ef94098f4e79e5251ffc5a770cda17e2be3/README.md","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/omni-/ask-jev/blob/74199ef94098f4e79e5251ffc5a770cda17e2be3/lib/Jev.psm1","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/omni-/ask-jev/blob/74199ef94098f4e79e5251ffc5a770cda17e2be3/lib/Report.psm1","note":"固定版本说明与源码；本站未执行该项目。"}],"pinned":false},{"id":"rhighs:jev-code","name":"jev-code","author":"rhighs","url":"https://github.com/rhighs/jev-code","category":"CLI & Pipelines","plainSummary":"一个实验性编程终端，让 Jev 逐步选择 AST 节点来组成 Python 或 Bash，也能作为命令行判断工具。","plainSummaryEn":"An experimental coding CLI where Jev selects AST productions for Python or Bash, with a separate command-line decision mode.","jevDecisionPoint":"从有限的语法与动作选项中选择，由本地程序生成代码或执行工具。","jevDecisionPointEn":"Chooses from bounded syntax and action options; local code constructs programs or invokes tools.","highlightBenefit":"把编程选择和命令行判断过程显示为可回看的记录。","highlightBenefitEn":"Shows coding choices and command-line judgments in inspectable traces.","tags":["ast-code-graph","cli-git-gates","typed-decisions"],"stars":19,"forks":2,"openIssues":0,"license":null,"language":"TypeScript","lastCommitAt":"2026-09-20T11:40:16Z","createdAt":"2026-09-17T17:18:16Z","metadataFetchedAt":"2026-09-21T21:56:59.941Z","avatarUrl":"https://avatars.githubusercontent.com/u/37136851?v=4","summarySource":"human-reviewed","claimStatus":"已核对固定版本的 README 与接入代码。","claimStatusEn":"Reviewed the README and integration code at a fixed commit. This site has not run the project or independently verified performance, security, or business outcomes.","plainSummaryJa":"Jev が AST の要素を選んで Python や Bash を組み立てる実験的 CLI。単独の判断コマンドも備える。","plainSummaryKo":"Jev가 AST 요소를 골라 Python이나 Bash를 구성하는 실험적 CLI로, 별도 판단 명령도 제공한다.","jevDecisionPointJa":"限られた構文や操作から選択し、ローカルコードがプログラム生成やツール呼び出しを行う。","jevDecisionPointKo":"제한된 구문과 동작에서 선택하면 로컬 코드가 프로그램을 구성하거나 도구를 호출한다.","highlightBenefitJa":"コードの選択とコマンドの判断を履歴で確認できる。","highlightBenefitKo":"코드 선택과 명령 판단 과정을 기록으로 확인할 수 있다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"unconfirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/rhighs/jev-code/blob/51bf5c1242a51feba8c5a4e18b61e379a9e94e5e/src/provider.ts","note":"Fixed-commit source reviewed / 已核对固定版本源码"},{"url":"https://github.com/rhighs/jev-code/blob/51bf5c1242a51feba8c5a4e18b61e379a9e94e5e/README.md","note":"Fixed-commit source reviewed / 已核对固定版本源码"}],"pinned":false},{"id":"lukaske:jev-doom-agent","name":"jev-doom-agent","author":"lukaske","url":"https://github.com/lukaske/jev-doom-agent","category":"High-Frequency & Simulation","plainSummary":"在浏览器里跑两份 Doom 引擎，让 Jev 根据游戏状态选择战术动作。","plainSummaryEn":"A browser Doom experiment comparing Jev-controlled players from the same initial state.","jevDecisionPoint":"读取结构化血量、弹药和可见目标，从战术宏中选一项，再交给本地控制器执行。","jevDecisionPointEn":"Jev reads structured health, ammunition and targets and selects tactical controls executed by a local controller.","highlightBenefit":"能在相同起点对照不同策略，并查看每次模型决策。","highlightBenefitEn":"Exposes model decisions alongside explicitly labeled offline and fallback policies.","tags":["games-simulation","typed-decisions"],"stars":7,"forks":0,"openIssues":0,"license":null,"language":"TypeScript","lastCommitAt":"2026-09-17T09:05:23Z","createdAt":"2026-09-17T08:37:37Z","metadataFetchedAt":"2026-09-21T21:57:02.931Z","avatarUrl":"https://avatars.githubusercontent.com/u/36110818?v=4","summarySource":"human-reviewed","claimStatus":"演示使用 Freedoom；Jev 接收结构化状态而非像素，失败路径有单独标记。","claimStatusEn":"The demo uses Freedoom. Jev receives structured state, not pixels, and fallback paths are labeled.","plainSummaryJa":"同じ初期状態から Jev が操作するプレイヤーを比較する、ブラウザー版 Doom 実験。","plainSummaryKo":"같은 초기 상태에서 Jev가 조종하는 플레이어를 비교하는 브라우저 Doom 실험.","jevDecisionPointJa":"体力・弾薬・対象の構造化情報から操作を選び、ローカル制御器が実行する。","jevDecisionPointKo":"체력·탄약·대상의 구조화 정보로 동작을 고르고 로컬 제어기가 실행한다.","highlightBenefitJa":"モデル判断と、明示されたオフライン・代替方針を見分けられる。","highlightBenefitKo":"모델 판단과 표시된 오프라인·대체 정책을 구분할 수 있다.","claimStatusJa":"固定バージョンの文書とソースを確認。実行・性能・安全性は当サイトで独立検証していません。","claimStatusKo":"고정 버전 문서와 소스를 확인했습니다. 실행·성능·안전성은 이 사이트에서 독립적으로 검증하지 않았습니다.","licenseStatus":"unconfirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/lukaske/jev-doom-agent/blob/318c32a24851444c1170bf083671c38723f3a35a/README.md","note":"README.md"},{"url":"https://github.com/lukaske/jev-doom-agent/blob/318c32a24851444c1170bf083671c38723f3a35a/server/typesafe.ts","note":"server/typesafe.ts"}],"pinned":false},{"id":"anessbelbati:jev-rerank-bench","name":"jev-rerank-bench","author":"anessbelbati","url":"https://github.com/anessbelbati/jev-rerank-bench","category":"Evaluation & Observability","plainSummary":"比较 Jev、专用 reranker 和聊天模型对同一批搜索片段的排序结果。","plainSummaryEn":"Compares Jev, dedicated rerankers and chat models on the same retrieved passages.","jevDecisionPoint":"用 Choice、Noul 和分档评分给候选片段排序，再计算检索指标。","jevDecisionPointEn":"Ranks candidate passages with Choice, Noul and rubric scores, then computes retrieval metrics.","highlightBenefit":"提供原始响应、评分代码和分数据集结果，便于检查比较条件。","highlightBenefitEn":"Publishes raw responses, scoring code and per-dataset results for inspection.","tags":["evaluation-benchmarks","search-retrieval","classification-ranking"],"stars":5,"forks":0,"openIssues":1,"license":"MIT","lastCommitAt":"2026-09-17T01:01:15Z","createdAt":"2026-09-16T20:43:41Z","metadataFetchedAt":"2026-09-21T21:57:05.901Z","avatarUrl":"https://avatars.githubusercontent.com/u/102622182?v=4","summarySource":"human-reviewed","claimStatus":"作者报告的八个英文数据集等权 nDCG@10 为 Jev 0.692、Cohere 0.691，未证明明确胜出；按查询等权排名会变化。 已核对固定版本 README 与集成源码。","claimStatusEn":"The author reports equal-dataset nDCG@10 of 0.692 for Jev and 0.691 for Cohere on eight English datasets, without establishing a winner. Query weighting changes the comparison. README and integration source reviewed at a fixed commit; not independently run or benchmarked by this site.","plainSummaryJa":"同じ検索候補で Jev、専用 reranker、チャットモデルの順位付けを比較する。","plainSummaryKo":"같은 검색 후보를 대상으로 Jev, 전용 reranker와 채팅 모델의 순위 결과를 비교한다.","jevDecisionPointJa":"Choice、Noul、段階評価で候補を並べ替え、検索指標を計算する。","jevDecisionPointKo":"Choice, Noul과 등급 점수로 후보를 정렬한 뒤 검색 지표를 계산한다.","highlightBenefitJa":"生の応答、評価コード、データセット別結果を公開している。","highlightBenefitKo":"원본 응답, 평가 코드와 데이터셋별 결과를 공개한다.","claimStatusJa":"固定版の README と実装を確認済み。このサイトでは独立した実行・性能検証をしていません。","claimStatusKo":"고정 버전 README와 구현을 확인했습니다. 이 사이트에서 독립적인 실행·성능 검증은 하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/anessbelbati/jev-rerank-bench/blob/cd9a35b22aeb4187334f7018a0ee1960a7470586/rerankers/jev.py#L28","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/anessbelbati/jev-rerank-bench/blob/cd9a35b22aeb4187334f7018a0ee1960a7470586/README.md","note":"固定版本的实现与说明 / Pinned source and documentation"}],"pinned":false},{"id":"danielkillenberger:jev-predict-skill","name":"jev-predict-skill","author":"DanielKillenberger","url":"https://github.com/DanielKillenberger/jev-predict-skill","category":"Decision Tools","plainSummary":"一个可供 Agent 执行的 skill 配方，根据规则和证据预测另一个 skill 的闭集结论。","plainSummaryEn":"An Agent skill recipe that predicts another skill’s closed-set outcome from its rules and evidence.","jevDecisionPoint":"先用 Jev 判断任务是否可拆成闭集问题，再从原 skill 的候选结论中选择。","jevDecisionPointEn":"Jev first assesses whether a closed decision is possible, then selects among the target skill’s outcomes.","highlightBenefit":"附带 API 调用与结果检查示例；它预测结论，不执行目标 skill。","highlightBenefitEn":"Includes API-call and response-check examples; it predicts the outcome without executing the target skill.","tags":["mcp-integrations","typed-decisions"],"stars":3,"forks":0,"openIssues":0,"license":null,"language":"HTML","lastCommitAt":"2026-09-16T21:41:44Z","createdAt":"2026-09-16T18:38:37Z","metadataFetchedAt":"2026-09-21T21:57:09.005Z","avatarUrl":"https://avatars.githubusercontent.com/u/5844780?v=4","summarySource":"human-reviewed","claimStatus":"核对了 skill 中的 API 调用和响应检查配方；未执行该 skill。","claimStatusEn":"Reviewed the API-call and response-validation recipe in the skill; the skill was not executed.","plainSummaryJa":"規則と証拠から別の skill の閉じた選択肢での結論を予測する Agent 用レシピ。","plainSummaryKo":"규칙과 증거로 다른 skill의 닫힌 선택지 결론을 예측하는 Agent용 레시피.","jevDecisionPointJa":"Jev が閉集合判断の可否を評価し、対象 skill の結論候補から選ぶ。","jevDecisionPointKo":"Jev가 닫힌 선택지 판단이 가능한지 평가한 뒤 대상 skill의 결론 후보를 선택한다.","highlightBenefitJa":"API 呼び出しと応答検査の例を含むが、対象 skill 自体は実行しない。","highlightBenefitKo":"API 호출·응답 검사 예제를 포함하며 대상 skill 자체를 실행하지 않는다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"unconfirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/DanielKillenberger/jev-predict-skill/blob/c80051db825de4a0af6efb03803e682daa466996/README.md","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/DanielKillenberger/jev-predict-skill/blob/c80051db825de4a0af6efb03803e682daa466996/SKILL.md","note":"固定版本说明与源码；本站未执行该项目。"}],"pinned":false},{"id":"aowang-ai:jev-trade","name":"jev-trade","author":"aowang-ai","url":"https://github.com/aowang-ai/jev-trade","category":"Domain & Vertical Tools","plainSummary":"一个 Hyperliquid 交易机器人实验，用 Jev 判断做多、做空，以及开仓、平仓或等待。","plainSummaryEn":"An experimental Hyperliquid trading bot using Jev to choose long or short and whether to open, close, or hold.","jevDecisionPoint":"每个资产账户读取行情后单独调用 Jev，执行器按选择提交或撤销订单。","jevDecisionPointEn":"Each asset account sends market state to Jev; execution code places or cancels orders based on the decisions.","highlightBenefit":"把模型决策、订单执行与看板状态分开记录。","highlightBenefitEn":"Separately records model decisions, order execution, and dashboard state.","tags":["finance-quant","typed-decisions"],"stars":88,"forks":19,"openIssues":1,"license":null,"language":"TypeScript","lastCommitAt":"2026-09-21T14:19:39Z","createdAt":"2026-09-17T16:33:05Z","metadataFetchedAt":"2026-09-21T21:57:11.898Z","avatarUrl":"https://avatars.githubusercontent.com/u/127380378?v=4","summarySource":"human-reviewed","claimStatus":"基于 jarrodwatts/jev-trader 改为 Hyperliquid 场所，存在实际改动；默认 mock 与 dry-run 不等于实盘。配置真实 key 可发送订单，本站未验证盈利或实盘表现。","claimStatusEn":"Adapts jarrodwatts/jev-trader to Hyperliquid with substantive venue changes. Default mock and dry-run modes are not live trading. Real keys can place orders; this site has not verified profitability or live performance.","plainSummaryJa":"Jev が売買方向と開始、決済、待機を選ぶ Hyperliquid 取引ボットの実験。","plainSummaryKo":"Jev가 매매 방향과 진입, 청산, 대기를 선택하는 Hyperliquid 거래 봇 실험.","jevDecisionPointJa":"各資産の口座が相場を Jev に送り、実行コードが注文や取り消しを行う。","jevDecisionPointKo":"각 자산 계정이 시장 상태를 Jev에 보내고 실행 코드가 주문하거나 취소한다.","highlightBenefitJa":"モデルの判断、注文実行、画面の状態を分けて記録する。","highlightBenefitKo":"모델 판단, 주문 실행, 대시보드 상태를 분리해 기록한다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"confirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/aowang-ai/jev-trade/blob/df2c9656324a8a75996eb0612de7adcfe3ce6f89/src/model.ts","note":"Fixed-commit source reviewed / 已核对固定版本源码"},{"url":"https://github.com/aowang-ai/jev-trade/blob/df2c9656324a8a75996eb0612de7adcfe3ce6f89/README.md","note":"Fixed-commit source reviewed / 已核对固定版本源码"},{"url":"https://github.com/aowang-ai/jev-trade/blob/df2c9656324a8a75996eb0612de7adcfe3ce6f89/LICENSE","note":"Fixed-commit source reviewed / 已核对固定版本源码"}],"pinned":false},{"id":"unownone:jevsume","name":"jevsume","author":"unownone","url":"https://github.com/unownone/jevsume","category":"Domain & Vertical Tools","plainSummary":"给简历做一次结构化体检。既检查措辞、结构和机器可读性，也能对照具体职位描述，看这份简历是否匹配。","plainSummaryEn":"A resume-review app that checks general writing and structure or compares a resume with a specific job description.","jevDecisionPoint":"对提取出的简历文本与职位要求逐项判断和打分，再由 Worker 汇总成页面里的评审结果。","jevDecisionPointEn":"Jev evaluates extracted resume text against typed questions; a Worker combines the results into review output.","highlightBenefit":"能保存输入、问题与输出，方便回看每次评审；没有 API key 时运行的是模拟结果。","highlightBenefitEn":"Records inputs and judgments so individual reviews can be inspected later.","tags":["domain-workflows","classification-ranking"],"stars":2,"forks":0,"openIssues":0,"license":null,"language":"TypeScript","lastCommitAt":"2026-09-21T21:06:53Z","createdAt":"2026-09-17T20:15:41Z","metadataFetchedAt":"2026-09-21T21:57:14.933Z","avatarUrl":"https://avatars.githubusercontent.com/u/38348296?v=4","summarySource":"human-reviewed","claimStatus":"无 API key 时使用模拟结果；简历评分不能证明招聘效果。","claimStatusEn":"Without an API key it uses mock results. Resume scores do not establish hiring outcomes.","plainSummaryJa":"履歴書の文章・構成を確認し、特定の求人との適合も調べるアプリ。","plainSummaryKo":"이력서의 문장·구조를 점검하고 특정 채용 공고와의 적합성도 평가하는 앱.","jevDecisionPointJa":"抽出した履歴書を Jev が質問ごとに評価し、Worker が結果をまとめる。","jevDecisionPointKo":"추출한 이력서를 Jev가 질문별로 평가하고 Worker가 결과를 모은다.","highlightBenefitJa":"入力と判断を記録し、個々のレビューを後で確認できる。","highlightBenefitKo":"입력과 판단을 기록해 개별 평가를 나중에 확인할 수 있다.","claimStatusJa":"固定バージョンの文書とソースを確認。実行・性能・安全性は当サイトで独立検証していません。","claimStatusKo":"고정 버전 문서와 소스를 확인했습니다. 실행·성능·안전성은 이 사이트에서 독립적으로 검증하지 않았습니다.","licenseStatus":"unconfirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/unownone/jevsume/blob/650a96a4c90da129cf745687d25abca36120c470/readme.md","note":"readme.md"},{"url":"https://github.com/unownone/jevsume/blob/650a96a4c90da129cf745687d25abca36120c470/packages/jev/index.ts","note":"packages/jev/index.ts"},{"url":"https://github.com/unownone/jevsume/blob/650a96a4c90da129cf745687d25abca36120c470/packages/jev/http.ts","note":"packages/jev/http.ts"}],"pinned":false},{"id":"thesammykins:jev_ampcode","name":"jev_ampcode","author":"thesammykins","url":"https://github.com/thesammykins/jev_ampcode","category":"MCP & Integrations","plainSummary":"给 Amp 的方案比较插件：只比较已经提供的选项、证据和偏好。","plainSummaryEn":"An Amp plugin for comparing supplied alternatives against supplied evidence and priorities.","jevDecisionPoint":"用 Jev Choice 比较有限候选，并返回概率和可供人工继续处理的结果。","jevDecisionPointEn":"Uses Jev Choice to compare bounded options and return probabilities for further review.","highlightBenefit":"让方案比较保留明确的候选与输入依据。","highlightBenefitEn":"Keeps alternatives and decision inputs explicit.","tags":["coding-agents","mcp-integrations","typed-decisions"],"stars":1,"forks":0,"openIssues":0,"license":null,"lastCommitAt":"2026-09-18T01:47:23Z","createdAt":"2026-09-18T01:13:21Z","metadataFetchedAt":"2026-09-21T21:57:18.027Z","avatarUrl":"https://avatars.githubusercontent.com/u/26049947?v=4","summarySource":"human-reviewed","claimStatus":"作者明确限定为咨询性 V1 工具，不负责研究、授权、正确性或安全保证。 已核对固定版本 README 与集成源码。","claimStatusEn":"The author limits this V1 tool to advisory comparisons. It does not supply research, authorization, correctness or safety guarantees. README and integration source reviewed at a fixed commit; not independently run or benchmarked by this site.","plainSummaryJa":"与えられた選択肢、根拠、優先事項を比較する Amp プラグイン。","plainSummaryKo":"제공된 선택지, 근거와 우선순위를 비교하는 Amp 플러그인이다.","jevDecisionPointJa":"Jev Choice で限定された候補を比較し、確認用の確率を返す。","jevDecisionPointKo":"Jev Choice로 한정된 후보를 비교하고 검토용 확률을 반환한다.","highlightBenefitJa":"候補と判断材料を明示できる。","highlightBenefitKo":"선택지와 판단 근거를 명확히 남긴다.","claimStatusJa":"固定版の README と実装を確認済み。このサイトでは独立した実行・性能検証をしていません。","claimStatusKo":"고정 버전 README와 구현을 확인했습니다. 이 사이트에서 독립적인 실행·성능 검증은 하지 않았습니다.","licenseStatus":"unconfirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/thesammykins/jev_ampcode/blob/8741655927f4ae8ea476f833795d3bdfa6e8f293/decision.ts#L74","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/thesammykins/jev_ampcode/blob/8741655927f4ae8ea476f833795d3bdfa6e8f293/README.md","note":"固定版本的实现与说明 / Pinned source and documentation"}],"pinned":false},{"id":"nicolasmontone:jev-tool-permissions","name":"jev-tool-permissions","author":"NicolasMontone","url":"https://github.com/NicolasMontone/jev-tool-permissions","category":"Security & Guardrails","plainSummary":"为 Vercel AI SDK 提供工具调用审批与工具列表筛选。","plainSummaryEn":"Adds tool-call approval and tool-list pruning to the Vercel AI SDK.","jevDecisionPoint":"确定性规则先处理；Jev 判断剩余调用风险和工具相关性，再由阈值映射结果。","jevDecisionPointEn":"Deterministic rules run first; Jev assesses remaining call risks and tool relevance for threshold-based handling.","highlightBenefit":"审批调用失败时要求人工确认；分类判断仍不能保证识别全部风险。","highlightBenefitEn":"Approval-call failures request human review; model judgments cannot guarantee detection of every risk.","tags":["security-guardrails","multilanguage-sdk","llm-routing-cost"],"stars":4,"forks":0,"openIssues":0,"license":null,"language":"TypeScript","lastCommitAt":"2026-09-18T00:45:29Z","createdAt":"2026-09-18T00:25:08Z","metadataFetchedAt":"2026-09-21T21:57:20.944Z","avatarUrl":"https://avatars.githubusercontent.com/u/83733486?v=4","summarySource":"human-reviewed","claimStatus":"已核对固定版本 README 与调用源码。","claimStatusEn":"Reviewed the fixed-version README and call path.","plainSummaryJa":"Vercel AI SDK にツール呼び出し承認とツール一覧の絞り込みを加える。","plainSummaryKo":"Vercel AI SDK에 도구 호출 승인과 도구 목록 필터링을 추가한다.","jevDecisionPointJa":"確定ルールを先に適用し、残りのリスクと関連性を Jev と閾値で処理する。","jevDecisionPointKo":"결정론적 규칙을 먼저 적용하고 남은 위험·관련성을 Jev와 임곗값으로 처리한다.","highlightBenefitJa":"承認 API の失敗時は人の確認へ回すが、全リスク検知を保証しない。","highlightBenefitKo":"승인 API 실패 시 사람의 확인을 요청하지만 모든 위험 탐지를 보장하지 않는다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"unconfirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/NicolasMontone/jev-tool-permissions/blob/4c57dd6b0251213353ca01ec74e2a44e04fa42e0/README.md","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/NicolasMontone/jev-tool-permissions/blob/4c57dd6b0251213353ca01ec74e2a44e04fa42e0/src/gate.ts","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/NicolasMontone/jev-tool-permissions/blob/4c57dd6b0251213353ca01ec74e2a44e04fa42e0/src/prune.ts","note":"固定版本说明与源码；本站未执行该项目。"}],"pinned":false},{"id":"cocktailpeanut:jevthoven","name":"jevthoven","author":"cocktailpeanut","url":"https://github.com/cocktailpeanut/jevthoven","category":"Creative Tools","plainSummary":"用一句话描述音乐，让 Jev 选择乐器、和声与逐小节片段，生成可编辑的多轨 MIDI。","plainSummaryEn":"Turns a music prompt into editable multitrack MIDI by asking Jev to choose instruments, harmony, and bar patterns.","jevDecisionPoint":"从曲式、乐器、和弦和节奏候选中逐步选择，本地程序把选择转为音符。","jevDecisionPointEn":"Selects form, instruments, chords, and rhythm candidates; local code renders the choices as notes.","highlightBenefit":"保留可编辑的音轨和决策记录，并支持播放与 MIDI 导出。","highlightBenefitEn":"Preserves editable tracks and decision receipts with playback and MIDI export.","tags":["creative-multimedia","typed-decisions"],"stars":10,"forks":3,"openIssues":1,"license":"MIT","language":"TypeScript","lastCommitAt":"2026-09-18T23:41:16Z","createdAt":"2026-09-17T21:25:55Z","metadataFetchedAt":"2026-09-21T21:57:23.922Z","avatarUrl":"https://avatars.githubusercontent.com/u/121128867?v=4","summarySource":"human-reviewed","claimStatus":"fixture 模式使用模拟决策，真实生成需要 TypeSafe key；这是符号音乐/MIDI 工具，不是音频生成模型。","claimStatusEn":"Fixture mode uses simulated decisions; live generation requires a TypeSafe key. This is a symbolic music/MIDI tool, not an audio-generation model.","plainSummaryJa":"音楽の説明から Jev が楽器、和声、小節パターンを選び、編集可能なマルチトラック MIDI を作る。","plainSummaryKo":"음악 설명을 바탕으로 Jev가 악기, 화성, 마디 패턴을 선택해 편집 가능한 멀티트랙 MIDI를 만든다.","jevDecisionPointJa":"曲構成、楽器、和音、リズムを候補から選び、コードが音符に変換する。","jevDecisionPointKo":"곡 구성, 악기, 화음, 리듬 후보를 선택하고 코드가 이를 음표로 변환한다.","highlightBenefitJa":"編集可能なトラックと判断記録を残し、再生と MIDI 書き出しに対応する。","highlightBenefitKo":"편집 가능한 트랙과 판단 기록을 보존하며 재생과 MIDI 내보내기를 지원한다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"confirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/cocktailpeanut/jevthoven/blob/e5e0c67d8bae92f96c9ce05138f55c3ab72c6aac/app/server/provider.ts","note":"Fixed-commit source reviewed / 已核对固定版本源码"},{"url":"https://github.com/cocktailpeanut/jevthoven/blob/e5e0c67d8bae92f96c9ce05138f55c3ab72c6aac/README.md","note":"Fixed-commit source reviewed / 已核对固定版本源码"}],"pinned":false},{"id":"amoghcreator:doom-jev","name":"doom-jev","author":"AmoghCreator","url":"https://github.com/AmoghCreator/doom-jev","category":"High-Frequency & Simulation","plainSummary":"让 Jev 玩 Doom：看结构化战况，决定往哪走、瞄谁和什么时候开火。","plainSummaryEn":"A ViZDoom Agent that uses Jev to choose movement, targets and firing from structured game state.","jevDecisionPoint":"选择宏观目标、敌人、移动、转向、跳跃与开火；几何逻辑接手细微瞄准。","jevDecisionPointEn":"Jev selects tactical goals and actions while local geometry handles fine aiming.","highlightBenefit":"把游戏运行和网络推理分开，不必每一帧都等待模型。","highlightBenefitEn":"Separates game ticks from network inference so the engine need not wait on every frame.","tags":["games-simulation","typed-decisions"],"stars":4,"forks":2,"openIssues":0,"license":null,"language":"Python","lastCommitAt":"2026-09-17T21:29:18Z","createdAt":"2026-09-17T21:11:21Z","metadataFetchedAt":"2026-09-21T21:57:26.939Z","avatarUrl":"https://avatars.githubusercontent.com/u/109089135?v=4","summarySource":"human-reviewed","claimStatus":"README 的推理频率和无掉帧表述是作者自述，本站未验证持续频率或通关效果。","claimStatusEn":"The README’s inference-rate and no-dropped-frame claims are author reports; sustained rates and game completion were not verified here.","plainSummaryJa":"構造化したゲーム状態から Jev が移動・標的・射撃を選ぶ ViZDoom Agent。","plainSummaryKo":"구조화된 게임 상태에서 Jev가 이동·표적·발사를 고르는 ViZDoom Agent.","jevDecisionPointJa":"Jev が戦術目標と動作を選び、細かな照準はローカルの幾何処理が担う。","jevDecisionPointKo":"Jev가 전술 목표와 동작을 고르고 세부 조준은 로컬 기하 계산이 맡는다.","highlightBenefitJa":"ゲームの更新とネットワーク推論を分離する。","highlightBenefitKo":"게임 갱신과 네트워크 추론을 분리한다.","claimStatusJa":"固定バージョンの文書とソースを確認。実行・性能・安全性は当サイトで独立検証していません。","claimStatusKo":"고정 버전 문서와 소스를 확인했습니다. 실행·성능·안전성은 이 사이트에서 독립적으로 검증하지 않았습니다.","licenseStatus":"unconfirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/AmoghCreator/doom-jev/blob/b27663fc0fa386f9d15ad6ea3c8f15374b855479/README.md","note":"README.md"},{"url":"https://github.com/AmoghCreator/doom-jev/blob/b27663fc0fa386f9d15ad6ea3c8f15374b855479/agent/jev_client.py","note":"agent/jev_client.py"}],"pinned":false},{"id":"zbush:jev-context","name":"jev-context","author":"zbush","url":"https://github.com/zbush/jev-context","category":"Context GC & Filter","plainSummary":"Codex 代码搜索插件：ripgrep 找候选，Jev 过滤后只返回判为相关的片段。","plainSummaryEn":"A Codex search plugin that filters ripgrep passages through Jev before returning relevant code.","jevDecisionPoint":"对代码片段与当前问题的相关性做判断，过滤 No 和 Unknown 结果。","jevDecisionPointEn":"Judges passage relevance and withholds No and Unknown results.","highlightBenefit":"记录过滤前后载荷，可按指定 tokenizer 核算返回文本的变化。","highlightBenefitEn":"Records paired payloads for measuring returned text under a named tokenizer.","tags":["context-compaction","coding-agents","search-retrieval"],"stars":1,"forks":0,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-19T17:22:17Z","createdAt":"2026-09-17T15:30:34Z","metadataFetchedAt":"2026-09-21T21:57:29.921Z","avatarUrl":"https://avatars.githubusercontent.com/u/5673410?v=4","summarySource":"human-reviewed","claimStatus":"过滤可能漏掉相关片段；返回文本 Token 节省不等于整项任务、上下文或账单节省。 已核对固定版本 README 与集成源码。","claimStatusEn":"Filtering can omit relevant passages. Returned-text token savings do not establish full-task, context-window or billing savings. README and integration source reviewed at a fixed commit; not independently run or benchmarked by this site.","plainSummaryJa":"ripgrep の候補を Jev で絞り、関連するコードを返す Codex 検索プラグイン。","plainSummaryKo":"ripgrep 후보를 Jev로 걸러 관련 코드를 반환하는 Codex 검색 플러그인이다.","jevDecisionPointJa":"質問との関連性を判断し、No と Unknown の候補を除く。","jevDecisionPointKo":"질문과의 관련성을 판단하고 No와 Unknown 결과를 제외한다.","highlightBenefitJa":"前後の内容を記録し、指定 tokenizer で返却量を比較できる。","highlightBenefitKo":"전후 내용을 기록해 지정 tokenizer로 반환량을 비교할 수 있다.","claimStatusJa":"固定版の README と実装を確認済み。このサイトでは独立した実行・性能検証をしていません。","claimStatusKo":"고정 버전 README와 구현을 확인했습니다. 이 사이트에서 독립적인 실행·성능 검증은 하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/zbush/jev-context/blob/2be700faebe862c259a16ac9f78266b728874ef6/src/jev.mjs#L46","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/zbush/jev-context/blob/2be700faebe862c259a16ac9f78266b728874ef6/README.md","note":"固定版本的实现与说明 / Pinned source and documentation"}],"pinned":false},{"id":"joevidev:ui-generator-instinct-jev","name":"ui-generator-instinct-jev","author":"joevidev","url":"https://github.com/joevidev/ui-generator-instinct-jev","category":"Creative Tools","plainSummary":"描述想要的界面，Jev 从既有 shadcn/ui 组件、字段和样式中选择，应用负责渲染。","plainSummaryEn":"Turns UI descriptions into selections from existing shadcn/ui components, fields and styles.","jevDecisionPoint":"将界面需求拆成选择和评分问题，结果映射到有限的组件目录。","jevDecisionPointEn":"Breaks requests into choice and score questions and maps answers to a bounded component catalog.","highlightBenefit":"演示用结构化选择组合 UI，Jev 本身不生成页面代码或文案。","highlightBenefitEn":"Demonstrates UI composition through decisions; Jev itself does not generate page code or copy.","tags":["creative-multimedia","typed-decisions"],"stars":6,"forks":0,"openIssues":0,"license":null,"language":"TypeScript","lastCommitAt":"2026-09-18T16:51:20Z","createdAt":"2026-09-17T04:08:39Z","metadataFetchedAt":"2026-09-21T21:57:32.955Z","avatarUrl":"https://avatars.githubusercontent.com/u/45459512?v=4","summarySource":"human-reviewed","claimStatus":"已核对固定版本 README 与调用源码。","claimStatusEn":"Reviewed the fixed-version README and call path.","plainSummaryJa":"UI の説明を既存の shadcn/ui コンポーネント・項目・スタイルの選択へ変換する。","plainSummaryKo":"UI 설명을 기존 shadcn/ui 컴포넌트·필드·스타일의 선택으로 변환한다.","jevDecisionPointJa":"要求を選択・採点問題に分け、回答を有限のコンポーネント一覧に対応させる。","jevDecisionPointKo":"요구를 선택·채점 질문으로 나누고 응답을 유한한 컴포넌트 목록에 연결한다.","highlightBenefitJa":"判断による UI 構成のデモで、Jev 自体はコードや文言を生成しない。","highlightBenefitKo":"판단으로 UI를 조합하는 데모이며 Jev 자체가 코드나 문구를 생성하지 않는다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"unconfirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/joevidev/ui-generator-instinct-jev/blob/9a81c2dc23b561980c7a6d4302c71f34e5ab4ae6/README.md","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/joevidev/ui-generator-instinct-jev/blob/9a81c2dc23b561980c7a6d4302c71f34e5ab4ae6/lib/decide.ts","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/joevidev/ui-generator-instinct-jev/blob/9a81c2dc23b561980c7a6d4302c71f34e5ab4ae6/lib/typesafe-client.ts","note":"固定版本说明与源码；本站未执行该项目。"}],"pinned":false},{"id":"droidrun:mobile-jev","name":"mobile-jev","author":"droidrun","url":"https://github.com/droidrun/mobile-jev","category":"Browser & OS Action","plainSummary":"通过 Mobilerun 控制 Android 手机，网页面板与命令行可查看 Jev 的操作过程。","plainSummaryEn":"Controls an Android phone through Mobilerun, with a web studio and CLI showing Jev decisions.","jevDecisionPoint":"根据手机界面选择应用、控件和下一步动作，Mobilerun 执行操作。","jevDecisionPointEn":"Chooses apps, controls and next actions from phone state; Mobilerun executes them.","highlightBenefit":"记录动作轨迹与请求时延，便于检查手机任务的执行过程。","highlightBenefitEn":"Records action traces and request latency for inspection.","tags":["desktop-os","typed-decisions"],"stars":322,"forks":41,"openIssues":4,"license":"MIT","lastCommitAt":"2026-09-17T12:51:55Z","createdAt":"2026-09-17T10:41:01Z","metadataFetchedAt":"2026-09-21T21:57:35.924Z","avatarUrl":"https://avatars.githubusercontent.com/u/207381488?v=4","summarySource":"human-reviewed","claimStatus":"作者演示约 21 秒完成九个动作并到付款选择页，不代表完成预约。 已核对固定版本 README 与集成源码。","claimStatusEn":"The author’s demo shows nine actions in about 21 seconds, ending at payment selection; it does not demonstrate a completed booking. README and integration source reviewed at a fixed commit; not independently run or benchmarked by this site.","plainSummaryJa":"Mobilerun 経由で Android を操作し、ウェブ画面と CLI で Jev の判断を確認できる。","plainSummaryKo":"Mobilerun으로 Android를 조작하며 웹 화면과 CLI에서 Jev 판단을 확인한다.","jevDecisionPointJa":"画面状態からアプリ、要素、次の操作を選び、Mobilerun が実行する。","jevDecisionPointKo":"화면 상태에서 앱, 요소와 다음 동작을 고르면 Mobilerun이 실행한다.","highlightBenefitJa":"操作履歴とリクエスト時間を記録する。","highlightBenefitKo":"동작 기록과 요청 지연을 남긴다.","claimStatusJa":"固定版の README と実装を確認済み。このサイトでは独立した実行・性能検証をしていません。","claimStatusKo":"고정 버전 README와 구현을 확인했습니다. 이 사이트에서 독립적인 실행·성능 검증은 하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/droidrun/mobile-jev/blob/395fc222beac4f059f9a0beb337d114a2b066e99/scripts/mobile-agent/policy.mjs#L224","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/droidrun/mobile-jev/blob/395fc222beac4f059f9a0beb337d114a2b066e99/README.md","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/droidrun/mobile-jev/blob/395fc222beac4f059f9a0beb337d114a2b066e99/scripts/mobile-agent/http.mjs","note":"固定版本的实现与说明 / Pinned source and documentation"}],"pinned":false},{"id":"minghanminghan:jev-demo","name":"jev-demo","author":"minghanminghan","url":"https://github.com/minghanminghan/jev-demo","category":"Routing & Cost Optimization","plainSummary":"Jev 客服分流演示：先批量回答路由问题，再沿分类结果处理用户请求。","plainSummaryEn":"A customer-service routing demo that batches Jev questions before following the resulting route.","jevDecisionPoint":"同一请求评估多层分类、转人工意愿与挫折程度，低置信度时交接。","jevDecisionPointEn":"Assesses classification levels, human-handoff intent and frustration, with low-confidence escalation.","highlightBenefit":"把分类和人工交接放在同一流程里，生成回复另由应用处理。","highlightBenefitEn":"Combines routing and human handoff while leaving response generation to the application.","tags":["domain-workflows","classification-ranking"],"stars":0,"forks":0,"openIssues":0,"license":null,"language":"TypeScript","lastCommitAt":"2026-09-17T22:32:13Z","createdAt":"2026-09-17T21:45:24Z","metadataFetchedAt":"2026-09-21T21:57:38.989Z","avatarUrl":"https://avatars.githubusercontent.com/u/122098568?v=4","summarySource":"human-reviewed","claimStatus":"已核对固定版本 README 与调用源码。","claimStatusEn":"Reviewed the fixed-version README and call path.","plainSummaryJa":"Jev にルート判断をまとめて問い、分類結果に沿って処理するカスタマーサービスのデモ。","plainSummaryKo":"Jev에 경로 질문을 묶어 보내고 분류 결과에 따라 처리하는 고객 지원 데모.","jevDecisionPointJa":"多段分類・人への引き継ぎ希望・不満度を評価し、低確信時にエスカレーションする。","jevDecisionPointKo":"다단계 분류·상담원 연결 의향·불만도를 평가하며 확신이 낮으면 넘긴다.","highlightBenefitJa":"分類と人への引き継ぎをまとめ、返信生成はアプリが担当する。","highlightBenefitKo":"분류와 상담원 연결을 묶고 응답 생성은 앱이 담당한다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"unconfirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/minghanminghan/jev-demo/blob/b6892294f97384bca86d68af9b56e0ea705d7893/README.md","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/minghanminghan/jev-demo/blob/b6892294f97384bca86d68af9b56e0ea705d7893/app/api/jev/[...path]/route.ts","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/minghanminghan/jev-demo/blob/b6892294f97384bca86d68af9b56e0ea705d7893/lib/oracle.ts","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/minghanminghan/jev-demo/blob/b6892294f97384bca86d68af9b56e0ea705d7893/lib/route-engine.ts","note":"固定版本说明与源码；本站未执行该项目。"}],"pinned":false},{"id":"adammichaelwood:jev-music-theory-1","name":"jev-music-theory-1","author":"adammichaelwood","url":"https://github.com/adammichaelwood/jev-music-theory-1","category":"Creative Tools","plainSummary":"用 Jev 做和声练习、乐理选择题，并通过连续选择和弦播放电钢琴。","plainSummaryEn":"Explores Jev on harmony exercises and music-theory questions, alongside a piano demo driven by chord choices.","jevDecisionPoint":"从声部、音高、时值或和弦候选中选择，代码评分或合成播放。","jevDecisionPointEn":"Chooses voices, pitches, durations, or chords; code grades exercises or synthesizes playback.","highlightBenefit":"把乐理测验和可听见的音乐实验放在同一项目中。","highlightBenefitEn":"Puts music-theory tests and an audible composition experiment in one project.","tags":["creative-multimedia","typed-decisions"],"stars":1,"forks":0,"openIssues":0,"license":null,"language":"TypeScript","lastCommitAt":"2026-09-18T03:20:59Z","createdAt":"2026-09-17T20:05:17Z","metadataFetchedAt":"2026-09-21T21:57:41.961Z","avatarUrl":"https://avatars.githubusercontent.com/u/1776241?v=4","summarySource":"human-reviewed","claimStatus":"作者报告了乐理关系推断失败，并披露实验由 AI 辅助完成；不能推广为通用乐理能力或独立基准。","claimStatusEn":"The author reports failures in relational music theory and discloses AI-assisted experimentation. Results should not be generalized into broad music-theory capability or an independent benchmark.","plainSummaryJa":"和声練習と楽理問題で Jev を試し、和音選択によるピアノ演奏も行う。","plainSummaryKo":"화성 연습과 음악 이론 문제로 Jev를 시험하고 화음 선택으로 피아노를 연주한다.","jevDecisionPointJa":"声部、音高、長さ、和音を選び、コードが採点または再生する。","jevDecisionPointKo":"성부, 음높이, 길이, 화음을 선택하면 코드가 채점하거나 재생한다.","highlightBenefitJa":"楽理のテストと音で確かめる実験を一つにまとめる。","highlightBenefitKo":"음악 이론 테스트와 직접 들을 수 있는 실험을 한 프로젝트에 담는다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"unconfirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/adammichaelwood/jev-music-theory-1/blob/7c5b4b8b7e40309f87388c3c5c4d3ed6a3872928/core/decide/jev.ts","note":"Fixed-commit source reviewed / 已核对固定版本源码"},{"url":"https://github.com/adammichaelwood/jev-music-theory-1/blob/7c5b4b8b7e40309f87388c3c5c4d3ed6a3872928/README.md","note":"Fixed-commit source reviewed / 已核对固定版本源码"}],"pinned":false},{"id":"jcpsimmons:jev-macos-loop","name":"jev-macos-loop","author":"jcpsimmons","url":"https://github.com/jcpsimmons/jev-macos-loop","category":"Browser & OS Action","plainSummary":"在 Mac 本地识别屏幕文字和控件，把文字选项交给 Jev，再点击它选中的元素。","plainSummaryEn":"A native macOS automation loop using local OCR and accessibility information with Jev action selection.","jevDecisionPoint":"Jev 从本地视觉、OCR 与无障碍标签组成的有限元素列表中选择；坐标和输入执行留在 Mac。","jevDecisionPointEn":"Jev selects controls or groups from locally observed candidates; coordinates and input execution stay on the Mac.","highlightBenefit":"让原生界面点击使用可枚举目标，并在执行前复核焦点与元素状态。","highlightBenefitEn":"Makes candidate controls and execution checks inspectable, including Finder workflows.","tags":["desktop-os","typed-decisions"],"stars":19,"forks":0,"openIssues":1,"license":"AGPL-3.0","language":"JavaScript","lastCommitAt":"2026-09-18T16:57:20Z","createdAt":"2026-09-18T05:17:42Z","metadataFetchedAt":"2026-09-21T21:57:44.896Z","avatarUrl":"https://avatars.githubusercontent.com/u/6600962?v=4","summarySource":"human-reviewed","claimStatus":"作者提供小规模 Finder 与设置演示；不能推广为任意桌面任务的可靠性保证。","claimStatusEn":"The author provides small Finder and settings demos, not broad reliability evidence for arbitrary desktop tasks; not retested here.","plainSummaryJa":"ローカル OCR とアクセシビリティ情報を使い、Jev が操作を選ぶ macOS 自動化ループ。","plainSummaryKo":"로컬 OCR과 접근성 정보를 사용하고 Jev가 동작을 고르는 macOS 자동화 루프.","jevDecisionPointJa":"観測した候補から対象を選び、座標処理と入力実行は Mac が担う。","jevDecisionPointKo":"관측한 후보에서 대상을 고르며 좌표 처리와 입력 실행은 Mac이 맡는다.","highlightBenefitJa":"Finder の操作を含め、候補と実行確認を追跡できる。","highlightBenefitKo":"Finder 작업을 포함해 후보와 실행 확인을 살펴볼 수 있다.","claimStatusJa":"固定バージョンの文書とソースを確認。実行・性能・安全性は当サイトで独立検証していません。","claimStatusKo":"고정 버전 문서와 소스를 확인했습니다. 실행·성능·안전성은 이 사이트에서 독립적으로 검증하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/jcpsimmons/jev-macos-loop/blob/1aadc01ef262c3b01460909c5dd102f0d9616ba5/README.md","note":"README.md"},{"url":"https://github.com/jcpsimmons/jev-macos-loop/blob/1aadc01ef262c3b01460909c5dd102f0d9616ba5/src/providers.mjs","note":"src/providers.mjs"}],"pinned":false,"topics":["accessibility","ai-agents","apple-silicon","apple-vision","computer-use","coreml","desktop-automation","gui-automation","jev","macos","omniparser","openrouter","swift","typesafeai","vercel-ai-gateway"]},{"id":"hugo-alves:jev-router-playground","name":"jev-router-playground","author":"hugo-alves","url":"https://github.com/hugo-alves/jev-router-playground","category":"Routing & Cost Optimization","plainSummary":"模型路由实验页：让 Jev 从候选模型中选择，再由你比较各模型的实际回答。","plainSummaryEn":"A model-routing playground where Jev picks a candidate and users compare the resulting answers.","jevDecisionPoint":"根据任务与候选模型说明选择模型，显示概率及运行记录。","jevDecisionPointEn":"Selects a model from the task and candidate profiles and records probabilities and runs.","highlightBenefit":"可导出选择结果，检查路由是否符合自己的回答偏好。","highlightBenefitEn":"Exports observations for checking routing choices against user preferences.","tags":["llm-routing-cost","evaluation-benchmarks"],"stars":2,"forks":0,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-18T01:25:53Z","createdAt":"2026-09-18T01:06:40Z","metadataFetchedAt":"2026-09-21T21:57:47.918Z","avatarUrl":"https://avatars.githubusercontent.com/u/3597209?v=4","summarySource":"human-reviewed","claimStatus":"默认 Jev 连接经过作者提供的 CORS 代理，不能把演示当成本地私密运行；本次未输入凭据或运行演示。 已核对固定版本 README 与集成源码。","claimStatusEn":"The default Jev connection passes through an author-provided CORS proxy. It should not be described as a private local-only execution path. No credentials were entered in this audit. README and integration source reviewed at a fixed commit; not independently run or benchmarked by this site.","plainSummaryJa":"Jev が候補モデルを選び、利用者が各回答を比較するルーティング実験画面。","plainSummaryKo":"Jev가 후보 모델을 고르고 사용자가 실제 답변을 비교하는 라우팅 실험 화면이다.","jevDecisionPointJa":"タスクと候補説明からモデルを選び、確率と実行を記録する。","jevDecisionPointKo":"작업과 후보 설명으로 모델을 선택하고 확률과 실행 기록을 남긴다.","highlightBenefitJa":"結果を出力し、自分の回答評価と選択を比較できる。","highlightBenefitKo":"기록을 내보내 자신의 답변 선호와 라우팅 결과를 비교한다.","claimStatusJa":"固定版の README と実装を確認済み。このサイトでは独立した実行・性能検証をしていません。","claimStatusKo":"고정 버전 README와 구현을 확인했습니다. 이 사이트에서 독립적인 실행·성능 검증은 하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/hugo-alves/jev-router-playground/blob/c2eb6ccf68b793d4ad99fb0ff1cc29edf11d9a93/app.js#L56","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/hugo-alves/jev-router-playground/blob/c2eb6ccf68b793d4ad99fb0ff1cc29edf11d9a93/README.md","note":"固定版本的实现与说明 / Pinned source and documentation"}],"pinned":false},{"id":"amberwhitehead:jevscript","name":"jevscript","author":"amberwhitehead","url":"https://github.com/amberwhitehead/jevscript","category":"CLI & Pipelines","plainSummary":"把语义判断作为语言原语的早期实验，目前实现的是 Jev 请求批处理验证脚本。","plainSummaryEn":"An early language experiment whose current implementation is a Jev request-batching spike.","jevDecisionPoint":"脚本比较单独与合并问题的回答、用量和延迟；完整语言引擎仍是设计目标。","jevDecisionPointEn":"The script compares separate and batched answers, usage and latency; the full language engine remains a design goal.","highlightBenefit":"适合研究请求批处理，不应当作已完成的编译器或解释器。","highlightBenefitEn":"Useful for studying batching, not a completed compiler or interpreter.","tags":["cli-git-gates","typed-decisions"],"stars":0,"forks":0,"openIssues":0,"license":null,"language":"JavaScript","lastCommitAt":"2026-09-17T02:36:37Z","createdAt":"2026-09-16T17:32:40Z","metadataFetchedAt":"2026-09-21T21:57:50.997Z","avatarUrl":"https://avatars.githubusercontent.com/u/131095?v=4","summarySource":"human-reviewed","claimStatus":"固定版本仅实现 M0 请求实验；完整语言引擎仍属规划，性能数字未独立复测。","claimStatusEn":"The checked version implements an M0 request experiment; the full language engine remains planned and performance was not independently retested.","plainSummaryJa":"意味判断を言語の原語にする初期実験。現状の実装は Jev 要求のバッチ化検証スクリプト。","plainSummaryKo":"의미 판단을 언어 기본 요소로 삼는 초기 실험으로 현재 구현은 Jev 요청 배치 검증 스크립트다.","jevDecisionPointJa":"個別・一括質問の回答、使用量、遅延を比較し、言語エンジン本体は設計段階。","jevDecisionPointKo":"개별·묶음 질문의 응답·사용량·지연을 비교하며 전체 언어 엔진은 설계 목표다.","highlightBenefitJa":"バッチ化の研究用で、完成したコンパイラーやインタープリターではない。","highlightBenefitKo":"배치 처리 연구용이며 완성된 컴파일러나 인터프리터는 아니다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"unconfirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/amberwhitehead/jevscript/blob/a01032218d8d57f84c5f310f585cd2da17671c9a/README.md","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/amberwhitehead/jevscript/blob/a01032218d8d57f84c5f310f585cd2da17671c9a/spike/m0.mjs","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/amberwhitehead/jevscript/blob/a01032218d8d57f84c5f310f585cd2da17671c9a/PLAN.md","note":"固定版本说明与源码；本站未执行该项目。"}],"pinned":false},{"id":"waxmell114514:jev-trade","name":"jev-trade","author":"Waxmell114514","url":"https://github.com/Waxmell114514/jev-trade","category":"Domain & Vertical Tools","plainSummary":"把 BTC、ETH 的行情变成文字状态，让 Jev 给交易判断，再放进含延迟和费用的模拟撮合里观察。","plainSummaryEn":"A simulated crypto trading loop that sends BTC and ETH market features to Jev and models execution costs and latency.","jevDecisionPoint":"代码先计算市场特征；Jev 回答有限的方向与风险问题，本地策略决定模拟仓位。","jevDecisionPointEn":"Jev answers bounded direction and risk questions; local policy decides simulated positions.","highlightBenefit":"把判断、延迟和执行成本放到同一实验记录里比较。","highlightBenefitEn":"Records judgments, delay and trading costs in the same experiment.","tags":["finance-quant","games-simulation"],"stars":1,"forks":0,"openIssues":0,"license":null,"language":"Python","lastCommitAt":"2026-09-21T13:04:22Z","createdAt":"2026-09-18T03:13:48Z","metadataFetchedAt":"2026-09-21T21:57:53.953Z","avatarUrl":"https://avatars.githubusercontent.com/u/232729665?v=4","summarySource":"human-reviewed","claimStatus":"模拟交易研究；回测或模拟盈亏不能证明未来盈利或实盘可用。","claimStatusEn":"A trading simulation study. Backtest or simulated P&L does not establish future profitability or live readiness.","plainSummaryJa":"BTC・ETH の特徴量を Jev に渡し、遅延と売買コストを含めて模擬取引するループ。","plainSummaryKo":"BTC·ETH 특징을 Jev에 보내고 지연과 거래 비용을 반영하는 모의거래 루프.","jevDecisionPointJa":"Jev が方向とリスクを判定し、模擬ポジションはローカル方針が決める。","jevDecisionPointKo":"Jev가 방향과 위험을 판단하고 로컬 정책이 모의 포지션을 결정한다.","highlightBenefitJa":"判断・遅延・コストを同じ実験記録で比較できる。","highlightBenefitKo":"판단·지연·비용을 같은 실험 기록에서 비교할 수 있다.","claimStatusJa":"固定バージョンの文書とソースを確認。実行・性能・安全性は当サイトで独立検証していません。","claimStatusKo":"고정 버전 문서와 소스를 확인했습니다. 실행·성능·안전성은 이 사이트에서 독립적으로 검증하지 않았습니다.","licenseStatus":"unconfirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/Waxmell114514/jev-trade/blob/df43b9636e34096f83457fe58bfe3f038879614c/README.md","note":"README.md"},{"url":"https://github.com/Waxmell114514/jev-trade/blob/df43b9636e34096f83457fe58bfe3f038879614c/jevtrade/jev/client.py","note":"jevtrade/jev/client.py"}],"pinned":false},{"id":"bud-ro:jev-demos","name":"jev-demos","author":"Bud-ro","url":"https://github.com/Bud-ro/jev-demos","category":"Evaluation & Observability","plainSummary":"用迷宫测试 Jev 的空间判断，比较单步选择和一次预测多步的表现。","plainSummaryEn":"Maze experiments comparing Jev single-step choices with multi-step lookahead.","jevDecisionPoint":"让 Jev 从方向候选中选择下一步或后续多步，再检查碰墙、绕路和到达情况。","jevDecisionPointEn":"Chooses moves from direction candidates, then checks collisions, paths and completion.","highlightBenefit":"保留实验设置和失败结果，便于研究判断边界。","highlightBenefitEn":"Keeps experiment settings and failures visible for studying limitations.","tags":["games-simulation","evaluation-benchmarks"],"stars":0,"forks":0,"openIssues":0,"license":null,"lastCommitAt":"2026-09-17T05:41:57Z","createdAt":"2026-09-17T01:35:25Z","metadataFetchedAt":"2026-09-21T21:57:56.949Z","avatarUrl":"https://avatars.githubusercontent.com/u/21696874?v=4","summarySource":"human-reviewed","claimStatus":"作者报告无邻接提示的 quick 设置未解出迷宫；不是已证明有效的实时导航引擎。 已核对固定版本 README 与集成源码。","claimStatusEn":"README and integration source reviewed at a fixed commit; not independently run or benchmarked by this site.","plainSummaryJa":"迷路で Jev の単一手選択と複数手の先読みを比較する実験。","plainSummaryKo":"미로에서 Jev의 한 단계 선택과 여러 단계 예측을 비교하는 실험이다.","jevDecisionPointJa":"方向候補から移動を選び、衝突、経路、到達を検査する。","jevDecisionPointKo":"방향 후보에서 움직임을 고른 뒤 충돌, 경로와 도착을 확인한다.","highlightBenefitJa":"実験条件と失敗を記録し、限界を調べられる。","highlightBenefitKo":"실험 조건과 실패를 남겨 판단 한계를 살펴볼 수 있다.","claimStatusJa":"固定版の README と実装を確認済み。このサイトでは独立した実行・性能検証をしていません。","claimStatusKo":"고정 버전 README와 구현을 확인했습니다. 이 사이트에서 독립적인 실행·성능 검증은 하지 않았습니다.","licenseStatus":"unconfirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/Bud-ro/jev-demos/blob/37470f96bc4cb91c5f10703caa7460bf0c20f23a/packages/jev_common/lib/src/client.dart#L232","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/Bud-ro/jev-demos/blob/37470f96bc4cb91c5f10703caa7460bf0c20f23a/packages/maze_lookahead/README.md","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/Bud-ro/jev-demos/blob/37470f96bc4cb91c5f10703caa7460bf0c20f23a/README.md","note":"固定版本的实现与说明 / Pinned source and documentation"}],"pinned":false},{"id":"jvsteiner:jevex","name":"jevex","author":"jvsteiner","url":"https://github.com/jvsteiner/jevex","category":"SDK & Decision Frameworks","plainSummary":"Jev 指挥工具循环的 Agent 实验，聊天模型负责参数与最终文字，MCP 工具执行操作。","plainSummaryEn":"An Agent experiment where Jev directs a tool loop, a chat model fills arguments and prose, and MCP tools execute.","jevDecisionPoint":"Jev 选择下一步动作并审批具体调用，执行结果回到下一轮状态。","jevDecisionPointEn":"Jev selects the next action and approves the concrete call; tool results feed the next state.","highlightBenefit":"把决策、内容生成和工具执行分开，附带作者的比较实验。","highlightBenefitEn":"Separates decisions, content generation and execution, with author-run comparisons.","tags":["mcp-integrations","typed-decisions"],"stars":3,"forks":0,"openIssues":1,"license":"MIT","language":"Python","lastCommitAt":"2026-09-17T12:53:19Z","createdAt":"2026-09-17T10:22:24Z","metadataFetchedAt":"2026-09-21T21:57:59.983Z","avatarUrl":"https://avatars.githubusercontent.com/u/1165328?v=4","summarySource":"human-reviewed","claimStatus":"已核对固定版本 README 与调用源码。","claimStatusEn":"Reviewed the fixed-version README and call path.","plainSummaryJa":"Jev がツール循環を指揮し、チャットモデルが引数・文章、MCP が実行を担う Agent 実験。","plainSummaryKo":"Jev가 도구 루프를 지휘하고 채팅 모델이 인수·문장을, MCP가 실행을 담당하는 Agent 실험.","jevDecisionPointJa":"次の動作と具体的な呼び出しを Jev が判断し、結果を次の状態に戻す。","jevDecisionPointKo":"Jev가 다음 동작과 구체적 호출을 판단하며 결과를 다음 상태에 반영한다.","highlightBenefitJa":"判断・生成・実行を分離し、作者による比較実験を含む。","highlightBenefitKo":"판단·생성·실행을 분리하며 저자의 비교 실험을 포함한다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/jvsteiner/jevex/blob/dd22ffd958d56ad118bb6dd585f1c913c96c5fd3/README.md","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/jvsteiner/jevex/blob/dd22ffd958d56ad118bb6dd585f1c913c96c5fd3/src/jevex/jev.py","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/jvsteiner/jevex/blob/dd22ffd958d56ad118bb6dd585f1c913c96c5fd3/src/jevex/agent.py","note":"固定版本说明与源码；本站未执行该项目。"}],"pinned":false},{"id":"mittal-parth:jev-experiments","name":"jev-experiments","author":"mittal-parth","url":"https://github.com/mittal-parth/jev-experiments","category":"High-Frequency & Simulation","plainSummary":"让 Jev 玩 Chrome 小恐龙和本地射击竞技场，Python 根据结构化判断执行动作。","plainSummaryEn":"Uses Jev to play Chrome Dino and a local shooter arena while Python executes structured decisions.","jevDecisionPoint":"从游戏状态中判断跳跃、蹲伏、移动、瞄准和开火，执行器应用本地规则。","jevDecisionPointEn":"Chooses jumping, ducking, movement, aim, or firing from game state, with local execution rules.","highlightBenefit":"把状态、模型答案和执行动作在检查界面中对照查看。","highlightBenefitEn":"An inspector lets users compare state, model answers, and executed actions.","tags":["games-simulation","typed-decisions"],"stars":0,"forks":0,"openIssues":0,"license":null,"language":"Python","lastCommitAt":"2026-09-18T05:19:16Z","createdAt":"2026-09-17T08:39:07Z","metadataFetchedAt":"2026-09-21T21:58:03.001Z","avatarUrl":"https://avatars.githubusercontent.com/u/76661350?v=4","summarySource":"human-reviewed","claimStatus":"项目包含明确标注的 heuristic fallback；模型读取结构化状态而非游戏截图，本站未验证运行频率或成绩。","claimStatusEn":"Includes an explicitly labeled heuristic fallback. The model reads structured state rather than game screenshots; this site has not verified decision rate or scores.","plainSummaryJa":"Jev が Chrome の恐竜ゲームやローカル FPS を判断し、Python が操作する。","plainSummaryKo":"Jev가 Chrome 공룡 게임과 로컬 FPS를 판단하고 Python이 동작을 실행한다.","jevDecisionPointJa":"状態からジャンプ、しゃがみ、移動、照準、射撃を選び、ローカルルールを適用する。","jevDecisionPointKo":"상태에서 점프, 숙이기, 이동, 조준, 발사를 고르고 로컬 규칙을 적용한다.","highlightBenefitJa":"検査画面で状態、回答、実行された操作を比較できる。","highlightBenefitKo":"검사 화면에서 상태, 답변, 실제 동작을 비교할 수 있다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"unconfirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/mittal-parth/jev-experiments/blob/2c8272ea13e2e4cee53f44dc8c6123b0c600edd6/dino_jev/client.py","note":"Fixed-commit source reviewed / 已核对固定版本源码"},{"url":"https://github.com/mittal-parth/jev-experiments/blob/2c8272ea13e2e4cee53f44dc8c6123b0c600edd6/README.md","note":"Fixed-commit source reviewed / 已核对固定版本源码"}],"pinned":false},{"id":"shifty-eye-games:foreman-jev","name":"foreman-jev","author":"Shifty-Eye-Games","url":"https://github.com/Shifty-Eye-Games/foreman-jev","category":"Evaluation & Observability","plainSummary":"给 Codex 工人配一个 Jev 监督员。它评估进展，但完成前还必须跑程序员指定的验收命令。","plainSummaryEn":"An experimental Jev supervisor for Codex workers with programmer-selected acceptance commands.","jevDecisionPoint":"对工作状态和证据做进展与完成判断，本地运行验收命令并检查源码在验证期间是否变化。","jevDecisionPointEn":"Jev assesses progress and completion signals; deterministic acceptance checks still decide whether the work passes.","highlightBenefit":"把模型判断和实际命令结果一并留档，方便检查任务是否真的收尾。","highlightBenefitEn":"Separates a model’s assessment from executable completion checks.","tags":["coding-agents","evaluation-benchmarks"],"stars":0,"forks":0,"openIssues":0,"license":"MIT","language":"Python","lastCommitAt":"2026-09-17T19:13:57Z","createdAt":"2026-09-17T19:16:52Z","metadataFetchedAt":"2026-09-21T21:58:05.974Z","avatarUrl":"https://avatars.githubusercontent.com/u/278721440?v=4","summarySource":"human-reviewed","claimStatus":"基于 foreman 的独立实验衍生版本；作者试跑证据不等于所有平台或任务都可靠，本站未重跑。","claimStatusEn":"An experimental derivative of foreman. The author’s pilot does not establish reliability across all platforms or tasks; not rerun here.","plainSummaryJa":"プログラマー指定の受入コマンドを備えた、Codex worker 用の Jev 監督実験。","plainSummaryKo":"개발자가 지정한 수락 명령을 갖춘 Codex worker용 Jev 감독 실험.","jevDecisionPointJa":"Jev が進捗や完了を評価し、最終合格は決定的な受入チェックで確認する。","jevDecisionPointKo":"Jev가 진행·완료 신호를 평가하며 최종 통과 여부는 결정적 수락 검사가 확인한다.","highlightBenefitJa":"モデルの評価と実行可能な完了確認を分離する。","highlightBenefitKo":"모델의 평가와 실행 가능한 완료 검사를 분리한다.","claimStatusJa":"固定バージョンの文書とソースを確認。実行・性能・安全性は当サイトで独立検証していません。","claimStatusKo":"고정 버전 문서와 소스를 확인했습니다. 실행·성능·안전성은 이 사이트에서 독립적으로 검증하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/Shifty-Eye-Games/foreman-jev/blob/3cb97e6051cb5764c50d3a0155f4d1483934fb25/README.md","note":"README.md"},{"url":"https://github.com/Shifty-Eye-Games/foreman-jev/blob/3cb97e6051cb5764c50d3a0155f4d1483934fb25/src/foreman/foreman/jev.py","note":"src/foreman/foreman/jev.py"}],"pinned":false},{"id":"reachjalil:jevlogs","name":"jevlogs","author":"reachjalil","url":"https://github.com/reachjalil/jevlogs","category":"Context GC & Filter","plainSummary":"在 OpenTelemetry 日志进入进一步分析前，用 Jev 标注诊断价值、优先级和路由信号。","plainSummaryEn":"Adds Jev diagnostic-value, priority and routing signals to OpenTelemetry logs before deeper analysis.","jevDecisionPoint":"按日志内容打分，判断是否值得送交更深入的模型分析。","jevDecisionPointEn":"Scores log records and judges whether they merit further model analysis.","highlightBenefit":"可把判断附在日志上，同时保留原归档路径。","highlightBenefitEn":"Can annotate records while preserving the existing archive path.","tags":["context-compaction","evaluation-benchmarks","classification-ranking"],"stars":9,"forks":1,"openIssues":3,"license":"MIT","lastCommitAt":"2026-09-17T07:29:42Z","createdAt":"2026-09-17T04:26:31Z","metadataFetchedAt":"2026-09-21T21:58:08.899Z","avatarUrl":"https://avatars.githubusercontent.com/u/6254931?v=4","summarySource":"human-reviewed","claimStatus":"已核对固定版本 README 与集成源码。","claimStatusEn":"README and integration source reviewed at a fixed commit; not independently run or benchmarked by this site.","plainSummaryJa":"OpenTelemetry ログに Jev の診断価値、優先度、振り分け判断を付ける。","plainSummaryKo":"OpenTelemetry 로그에 Jev의 진단 가치, 우선순위와 라우팅 판단을 붙인다.","jevDecisionPointJa":"各ログを評価し、追加のモデル解析に回す価値があるか判断する。","jevDecisionPointKo":"로그를 평가해 추가 모델 분석이 필요한지 판단한다.","highlightBenefitJa":"既存の保存経路を維持しながら判断を追記できる。","highlightBenefitKo":"기존 보관 경로를 유지하면서 판단을 추가할 수 있다.","claimStatusJa":"固定版の README と実装を確認済み。このサイトでは独立した実行・性能検証をしていません。","claimStatusKo":"고정 버전 README와 구현을 확인했습니다. 이 사이트에서 독립적인 실행·성능 검증은 하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/reachjalil/jevlogs/blob/b1ff60079c30d50a7f93dbfa09848f9539665d44/src/index.ts#L52","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/reachjalil/jevlogs/blob/b1ff60079c30d50a7f93dbfa09848f9539665d44/README.md","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/reachjalil/jevlogs/blob/b1ff60079c30d50a7f93dbfa09848f9539665d44/examples/receiver/README.md","note":"固定版本的实现与说明 / Pinned source and documentation"}],"pinned":false},{"id":"luxus:ha-conversation-jev","name":"ha-conversation-jev","author":"luxus","url":"https://github.com/luxus/ha-conversation-jev","category":"Voice & Conversation","plainSummary":"Home Assistant 的对话扩展：简单灯光指令走设备服务，其余请求交给 Grok。","plainSummaryEn":"A Home Assistant conversation integration routing simple lighting commands to services and other requests to Grok.","jevDecisionPoint":"Jev 对语句和已公开设备做分类，程序据此选灯光服务或对话分支。","jevDecisionPointEn":"Jev classifies the utterance and exposed devices; code chooses a lighting service or conversation path.","highlightBenefit":"为有限的灯光操作提供独立路径，不代表支持所有家居设备命令。","highlightBenefitEn":"Provides a bounded lighting fast path rather than support for every smart-home command.","tags":["domain-workflows","voice-conversation","llm-routing-cost"],"stars":2,"forks":1,"openIssues":0,"license":null,"language":"Python","lastCommitAt":"2026-09-19T20:40:24Z","createdAt":"2026-09-17T17:56:27Z","metadataFetchedAt":"2026-09-21T21:58:11.968Z","avatarUrl":"https://avatars.githubusercontent.com/u/7449?v=4","summarySource":"human-reviewed","claimStatus":"已核对固定版本 README 与调用源码。","claimStatusEn":"Reviewed the fixed-version README and call path.","plainSummaryJa":"簡単な照明指示をデバイスサービスへ、それ以外を Grok へ渡す Home Assistant の会話拡張。","plainSummaryKo":"간단한 조명 명령은 기기 서비스로, 나머지는 Grok으로 보내는 Home Assistant 대화 확장.","jevDecisionPointJa":"Jev が発話と公開デバイスを分類し、コードが照明操作か会話かを選ぶ。","jevDecisionPointKo":"Jev가 발화와 공개된 기기를 분류하면 코드가 조명 동작 또는 대화 경로를 선택한다.","highlightBenefitJa":"対応範囲は限定された照明操作で、全家電命令への対応ではない。","highlightBenefitKo":"제한된 조명 동작 경로이며 모든 스마트홈 명령을 지원한다는 뜻은 아니다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"unconfirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/luxus/ha-conversation-jev/blob/a405366b66b8c8dc5d73b32333043befdbbe2287/README.md","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/luxus/ha-conversation-jev/blob/a405366b66b8c8dc5d73b32333043befdbbe2287/custom_components/jev_assist/jev_client.py","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/luxus/ha-conversation-jev/blob/a405366b66b8c8dc5d73b32333043befdbbe2287/custom_components/jev_assist/jev_router.py","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/luxus/ha-conversation-jev/blob/a405366b66b8c8dc5d73b32333043befdbbe2287/custom_components/jev_assist/conversation.py","note":"固定版本说明与源码；本站未执行该项目。"}],"pinned":false},{"id":"vercel:ai","name":"ai","author":"vercel","url":"https://github.com/vercel/ai","category":"SDK & Decision Frameworks","plainSummary":"AI SDK 中的 TypeSafe provider，让 TypeScript 应用通过统一 evaluate 接口调用 Jev。","plainSummaryEn":"The TypeSafe provider in AI SDK lets TypeScript applications call Jev through the shared evaluate interface.","jevDecisionPoint":"把选择、评分和是非问题转换为 TypeSafe System One 请求，并解析结构化结果。","jevDecisionPointEn":"Maps choice, score, and yes/no questions to TypeSafe System One requests and parses typed results.","highlightBenefit":"在 AI SDK 应用中复用统一的调用接口。","highlightBenefitEn":"Reuses a common evaluation interface within AI SDK applications.","tags":["multilanguage-sdk","typed-decisions"],"stars":26877,"forks":5168,"openIssues":1475,"license":null,"language":"TypeScript","lastCommitAt":"2026-09-21T21:44:14Z","createdAt":"2023-05-23T15:04:08Z","metadataFetchedAt":"2026-09-21T21:58:14.971Z","avatarUrl":"https://avatars.githubusercontent.com/u/14985020?v=4","summarySource":"human-reviewed","claimStatus":"核实的是仓库内一个可选 provider，不代表整个 AI SDK 由 Jev 驱动或所有用户都在使用 Jev。","claimStatusEn":"The verified scope is one optional provider in the repository, not that all of AI SDK is powered by Jev or that all users use it.","plainSummaryJa":"AI SDK の TypeSafe provider から、共通の evaluate インターフェースで Jev を呼び出す。","plainSummaryKo":"AI SDK의 TypeSafe provider로 공통 evaluate 인터페이스에서 Jev를 호출한다.","jevDecisionPointJa":"選択、採点、二択の質問を System One リクエストに変換し、型付きの結果を解析する。","jevDecisionPointKo":"선택, 평가, 예·아니요 질문을 System One 요청으로 바꾸고 타입이 있는 결과를 해석한다.","highlightBenefitJa":"AI SDK アプリ内で共通の評価インターフェースを使える。","highlightBenefitKo":"AI SDK 애플리케이션에서 공통 평가 인터페이스를 재사용한다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"confirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/vercel/ai/blob/73ec7015edd4f04ca9144ce93a8a037a731e5db8/packages/typesafe-ai/src/typesafe-ai-evaluation-model.ts","note":"Fixed-commit source reviewed / 已核对固定版本源码"},{"url":"https://github.com/vercel/ai/blob/73ec7015edd4f04ca9144ce93a8a037a731e5db8/README.md","note":"Fixed-commit source reviewed / 已核对固定版本源码"},{"url":"https://github.com/vercel/ai/blob/73ec7015edd4f04ca9144ce93a8a037a731e5db8/LICENSE","note":"Fixed-commit source reviewed / 已核对固定版本源码"}],"pinned":false,"topics":["anthropic","artificial-intelligence","gemini","generative-ai","generative-ui","javascript","language-model","llm","nextjs","openai","react","svelte","typescript","vercel","vue"]},{"id":"langchain-ai:langchain","name":"langchain","author":"langchain-ai","url":"https://github.com/langchain-ai/langchain","category":"SDK & Integrations","plainSummary":"给 Python LangChain 流程加一个可选 Jev 分类节点，返回类别、概率和等级评分。","plainSummaryEn":"An optional Jev classifier integration for Python LangChain workflows.","jevDecisionPoint":"TypeSafeClassifier 将 JSON 状态及类型化问题发送到 /v1/systemone，支持同步与异步 Runnable 调用。","jevDecisionPointEn":"Submits binary, categorical and ordered-score questions and returns typed answers with probabilities.","highlightBenefit":"把结构化判断接进已有 LangChain 流程与追踪接口。","highlightBenefitEn":"Adds a classification step that can participate in existing LangChain pipelines.","tags":["multilanguage-sdk","classification-ranking"],"stars":146822,"forks":24552,"openIssues":542,"license":"MIT","language":"Python","lastCommitAt":"2026-09-21T18:23:30Z","createdAt":"2022-10-17T02:58:36Z","metadataFetchedAt":"2026-09-21T21:58:17.954Z","avatarUrl":"https://avatars.githubusercontent.com/u/126733545?v=4","summarySource":"human-reviewed","claimStatus":"核对的是仓库内 TypeSafe partner 模块，不表示整个 LangChain 默认使用 Jev。","claimStatusEn":"This entry covers the TypeSafe partner module, not a claim that LangChain uses Jev by default.","plainSummaryJa":"Python LangChain のワークフローに追加する、任意の Jev 分類連携。","plainSummaryKo":"Python LangChain 작업 흐름에 추가하는 선택형 Jev 분류 연동.","jevDecisionPointJa":"二値・分類・段階評価の質問を送り、型付き回答と確率を返す。","jevDecisionPointKo":"이진·범주·단계 평가 질문을 보내 타입이 있는 답과 확률을 받는다.","highlightBenefitJa":"既存 LangChain パイプラインへ分類処理を追加できる。","highlightBenefitKo":"기존 LangChain 파이프라인에 분류 단계를 추가한다.","claimStatusJa":"固定バージョンの文書とソースを確認。実行・性能・安全性は当サイトで独立検証していません。","claimStatusKo":"고정 버전 문서와 소스를 확인했습니다. 실행·성능·안전성은 이 사이트에서 독립적으로 검증하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/langchain-ai/langchain/blob/eba445b7563d1709427bd8072892975a6ea59fdc/README.md","note":"README.md"},{"url":"https://github.com/langchain-ai/langchain/blob/eba445b7563d1709427bd8072892975a6ea59fdc/libs/partners/typesafe/langchain_typesafe/classifier.py","note":"libs/partners/typesafe/langchain_typesafe/classifier.py"}],"pinned":false,"topics":["agents","ai","ai-agents","anthropic","chatgpt","deepagents","enterprise","framework","gemini","generative-ai","langchain","langgraph","llm","multiagent","open-source","openai","pydantic","python","rag","typescript"]},{"id":"can1357:oh-my-pi","name":"oh-my-pi","author":"can1357","url":"https://github.com/can1357/oh-my-pi","category":"Routing & Cost Optimization","plainSummary":"Oh My Pi 编程 Agent 内含可选的 TypeSafe 判断提供器，供小型决策流程调用。","plainSummaryEn":"Oh My Pi includes an optional TypeSafe judgment provider for bounded decisions in coding-agent workflows.","jevDecisionPoint":"将 Agent 的状态与有界问题发送给 Jev，并解析结构化回答。","jevDecisionPointEn":"Sends agent state and typed questions to Jev and parses structured answers.","highlightBenefit":"在现有 Agent 流程里接入可替换的判断提供器。","highlightBenefitEn":"Adds a replaceable judgment provider to existing agent workflows.","tags":["coding-agents","multilanguage-sdk","typed-decisions"],"stars":32299,"forks":3430,"openIssues":3042,"license":"MIT","lastCommitAt":"2026-09-21T21:23:14Z","createdAt":"2025-12-31T14:01:28Z","metadataFetchedAt":"2026-09-21T21:58:20.920Z","avatarUrl":"https://avatars.githubusercontent.com/u/11807264?v=4","summarySource":"human-reviewed","claimStatus":"已核对固定版本 README 与集成源码。","claimStatusEn":"README and integration source reviewed at a fixed commit; not independently run or benchmarked by this site.","plainSummaryJa":"Oh My Pi に含まれる任意の TypeSafe 判断プロバイダーを、Agent の限定的な判断で使える。","plainSummaryKo":"Oh My Pi에는 코딩 Agent의 한정된 판단에 쓰는 선택형 TypeSafe 제공자가 포함된다.","jevDecisionPointJa":"Agent の状態と型付き質問を Jev に送り、回答を解析する。","jevDecisionPointKo":"Agent 상태와 구조화된 질문을 Jev에 보내 응답을 처리한다.","highlightBenefitJa":"既存の Agent 処理に交換可能な判断プロバイダーを加えられる。","highlightBenefitKo":"기존 Agent 흐름에 교체 가능한 판단 제공자를 더한다.","claimStatusJa":"固定版の README と実装を確認済み。このサイトでは独立した実行・性能検証をしていません。","claimStatusKo":"고정 버전 README와 구현을 확인했습니다. 이 사이트에서 독립적인 실행·성능 검증은 하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/can1357/oh-my-pi/blob/78b753124d11f8dd3ae73e2524125890ff7c977e/packages/ai/src/judgment/typesafe.ts#L4","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/can1357/oh-my-pi/blob/78b753124d11f8dd3ae73e2524125890ff7c977e/README.md","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/can1357/oh-my-pi/blob/78b753124d11f8dd3ae73e2524125890ff7c977e/infra/docs/README.md","note":"固定版本的实现与说明 / Pinned source and documentation"}],"pinned":false},{"id":"elizaos:eliza","name":"eliza","author":"elizaOS","url":"https://github.com/elizaOS/eliza","category":"SDK & Decision Frameworks","plainSummary":"Eliza 源码中的可选 TypeSafe HTTP 适配器，默认没有注册到 Agent 运行时。","plainSummaryEn":"An optional TypeSafe HTTP adapter in Eliza’s source, not registered with the Agent runtime by default.","jevDecisionPoint":"只有业务代码显式调用 systemOne 才发送状态和问题，返回经校验的结构化答案。","jevDecisionPointEn":"Only an explicit systemOne call sends state and questions, returning validated typed answers.","highlightBenefit":"提供一个可复用的服务端接入模块，不代表 Eliza 已在实际业务中使用 Jev。","highlightBenefitEn":"Provides a reusable server-side module; its presence is not evidence of deployed Jev adoption.","tags":["multilanguage-sdk","typed-decisions"],"stars":19407,"forks":5750,"openIssues":1490,"license":"MIT","language":"TypeScript","lastCommitAt":"2026-09-21T20:28:30Z","createdAt":"2024-07-09T07:55:40Z","metadataFetchedAt":"2026-09-21T21:58:23.928Z","avatarUrl":"https://avatars.githubusercontent.com/u/186240462?v=4","summarySource":"human-reviewed","claimStatus":"适配器默认未注册运行时；源码存在不等于真实采用或已上线。","claimStatusEn":"The adapter is not registered with the runtime by default; source presence does not establish real adoption or deployment.","plainSummaryJa":"Eliza のソースにある任意の TypeSafe HTTP アダプター。既定では Agent 実行系に未登録。","plainSummaryKo":"Eliza 소스의 선택적 TypeSafe HTTP 어댑터로 기본 Agent 런타임에는 등록되지 않는다.","jevDecisionPointJa":"明示的な systemOne 呼び出しだけが状態と質問を送り、検証済みの型付き回答を返す。","jevDecisionPointKo":"명시적 systemOne 호출만 상태와 질문을 보내 검증된 구조화 응답을 반환한다.","highlightBenefitJa":"再利用可能なサーバーモジュールであり、実運用での Jev 採用を示すものではない。","highlightBenefitKo":"재사용할 서버 모듈이며 실제 서비스에서 Jev를 채택했다는 증거는 아니다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/elizaOS/eliza/blob/ebc808e3a67fb941e29153d89fc896524d32fe3c/README.md","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/elizaOS/eliza/blob/ebc808e3a67fb941e29153d89fc896524d32fe3c/packages/agent/src/services/typesafe/README.md","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/elizaOS/eliza/blob/ebc808e3a67fb941e29153d89fc896524d32fe3c/packages/agent/src/services/typesafe/client.ts","note":"固定版本说明与源码；本站未执行该项目。"}],"pinned":false,"topics":["agent","agentic","ai","autonomous","chatbot","crypto","discord","eliza","elizaos","framework","plugins","rag","slack","swarm","telegram"]},{"id":"composiohq:composio","name":"composio","author":"ComposioHQ","url":"https://github.com/ComposioHQ/composio","category":"SDK & Decision Frameworks","plainSummary":"Composio 的可选 TypeSafe provider，用 Jev 从工具与有限参数选项中做判断。","plainSummaryEn":"An optional TypeSafe provider for Composio that uses Jev to choose among tools and bounded argument options.","jevDecisionPoint":"将工具或操作条件转成结构化问题，读取 Jev 答案后交由本地调用逻辑处理。","jevDecisionPointEn":"Turns tool or action conditions into structured questions and passes Jev answers to local invocation logic.","highlightBenefit":"把结构化判断接到已有工具调用接口中。","highlightBenefitEn":"Adds structured judgments to existing tool interfaces.","tags":["multilanguage-sdk","llm-routing-cost","typed-decisions"],"stars":30275,"forks":4811,"openIssues":83,"license":"MIT","language":"TypeScript","lastCommitAt":"2026-09-21T20:57:11Z","createdAt":"2024-02-23T13:58:27Z","metadataFetchedAt":"2026-09-21T21:58:26.959Z","avatarUrl":"https://avatars.githubusercontent.com/u/128464815?v=4","summarySource":"human-reviewed","claimStatus":"核实的是可选 provider 接入；开放文本参数、执行策略和外部服务授权仍由调用方负责。","claimStatusEn":"The verified scope is an optional provider. Open-text arguments, execution policy, and external-service authorization remain the caller’s responsibility.","plainSummaryJa":"Composio の任意の TypeSafe provider が、ツールや限られた引数候補を Jev に判断させる。","plainSummaryKo":"Composio의 선택적 TypeSafe provider가 도구와 제한된 인자 후보를 Jev로 판단한다.","jevDecisionPointJa":"ツールや操作条件を構造化質問にし、回答をローカルの呼び出し処理に渡す。","jevDecisionPointKo":"도구나 작업 조건을 구조화된 질문으로 만들고 답변을 로컬 호출 로직에 전달한다.","highlightBenefitJa":"既存のツールインターフェースに構造化判断を加えられる。","highlightBenefitKo":"기존 도구 인터페이스에 구조화된 판단을 추가할 수 있다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"confirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/ComposioHQ/composio/blob/4b5920bf7aa55c8a44657b060d4bd25ce7b13a9a/ts/packages/providers/typesafe/src/decide.ts","note":"Fixed-commit source reviewed / 已核对固定版本源码"},{"url":"https://github.com/ComposioHQ/composio/blob/4b5920bf7aa55c8a44657b060d4bd25ce7b13a9a/README.md","note":"Fixed-commit source reviewed / 已核对固定版本源码"}],"pinned":false,"topics":["agentic-ai","agents","ai","ai-agents","aiagents","developer-tools","function-calling","gpt-4","javascript","js","llm","llmops","mcp","python","remote-mcp-server","sse","typescript"]},{"id":"pydantic:pydantic-ai","name":"pydantic-ai","author":"pydantic","url":"https://github.com/pydantic/pydantic-ai","category":"SDK & Integrations","plainSummary":"Pydantic AI 的可选 Jev 模型：把输出模型里的布尔和枚举字段变成问题，拿回符合类型的判断。","plainSummaryEn":"An optional TypeSafe provider and Jev model integration for Pydantic AI.","jevDecisionPoint":"TypeSafeModel 将支持的 output_type 字段编译为类型化问题，经 TypeSafeProvider 请求 API 后还原输出。","jevDecisionPointEn":"Converts supported structured output fields into typed Jev questions and maps answers back to the output model.","highlightBenefit":"决策型 Agent 可以复用 Pydantic 输出定义，并与其他模型作对照。","highlightBenefitEn":"Lets supported classification outputs use the existing Pydantic AI interface.","tags":["multilanguage-sdk","typed-decisions"],"stars":20103,"forks":2750,"openIssues":942,"license":"MIT","language":"Python","lastCommitAt":"2026-09-21T21:57:22Z","createdAt":"2024-06-21T15:55:04Z","metadataFetchedAt":"2026-09-21T21:58:29.997Z","avatarUrl":"https://avatars.githubusercontent.com/u/110818415?v=4","summarySource":"human-reviewed","claimStatus":"可选 provider，不是通用聊天模型替换；可用输出类型应以该版本实现为准。","claimStatusEn":"An optional provider rather than a general chat-model replacement; supported output types depend on this implementation. Not run here.","plainSummaryJa":"Pydantic AI 向けの任意の TypeSafe provider と Jev モデル連携。","plainSummaryKo":"Pydantic AI용 선택형 TypeSafe provider와 Jev 모델 연동.","jevDecisionPointJa":"対応する構造化出力のフィールドを Jev の質問に変換し、回答を出力モデルへ戻す。","jevDecisionPointKo":"지원하는 구조화 출력 필드를 Jev 질문으로 변환하고 답을 출력 모델에 매핑한다.","highlightBenefitJa":"対応する分類出力で Pydantic AI の既存インターフェースを使える。","highlightBenefitKo":"지원하는 분류 출력에 기존 Pydantic AI 인터페이스를 사용할 수 있다.","claimStatusJa":"固定バージョンの文書とソースを確認。実行・性能・安全性は当サイトで独立検証していません。","claimStatusKo":"고정 버전 문서와 소스를 확인했습니다. 실행·성능·안전성은 이 사이트에서 독립적으로 검증하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/pydantic/pydantic-ai/blob/c4898abb54dc25ae6f6aef208a4c0661b30a455e/README.md","note":"README.md"},{"url":"https://github.com/pydantic/pydantic-ai/blob/c4898abb54dc25ae6f6aef208a4c0661b30a455e/pydantic_ai_slim/pydantic_ai/providers/typesafe.py","note":"pydantic_ai_slim/pydantic_ai/providers/typesafe.py"},{"url":"https://github.com/pydantic/pydantic-ai/blob/c4898abb54dc25ae6f6aef208a4c0661b30a455e/pydantic_ai_slim/pydantic_ai/models/typesafe.py","note":"pydantic_ai_slim/pydantic_ai/models/typesafe.py"}],"pinned":false,"topics":["agent-framework","genai","harness","harness-engineering","llm","pydantic","python"]},{"id":"berriai:litellm","name":"litellm","author":"BerriAI","url":"https://github.com/BerriAI/litellm","category":"Routing & Cost Optimization","plainSummary":"LiteLLM 的复杂度路由器可选用 Jev 判断请求应交给哪个模型档位。","plainSummaryEn":"LiteLLM can use Jev to classify requests for its complexity-based model router.","jevDecisionPoint":"将请求映射到预设复杂度类别，再由 LiteLLM 路由策略选择后端。","jevDecisionPointEn":"Maps requests to configured complexity classes that drive backend routing.","highlightBenefit":"让模型分流使用可检查的复杂度判断。","highlightBenefitEn":"Makes the complexity judgment behind routing inspectable.","tags":["llm-routing-cost","multilanguage-sdk"],"stars":59343,"forks":11648,"openIssues":5231,"license":null,"lastCommitAt":"2026-09-21T21:51:53Z","createdAt":"2023-07-27T00:09:52Z","metadataFetchedAt":"2026-09-21T21:58:32.955Z","avatarUrl":"https://avatars.githubusercontent.com/u/121462774?v=4","summarySource":"human-reviewed","claimStatus":"已核对固定版本 README 与集成源码。","claimStatusEn":"README and integration source reviewed at a fixed commit; not independently run or benchmarked by this site.","plainSummaryJa":"LiteLLM の複雑度ルーターで、Jev にリクエストの分類を任せられる。","plainSummaryKo":"LiteLLM의 복잡도 라우터에서 Jev로 요청을 분류할 수 있다.","jevDecisionPointJa":"設定済みの複雑度へ分類し、その結果からバックエンドを振り分ける。","jevDecisionPointKo":"설정된 복잡도 등급에 매핑하고 그 결과로 백엔드를 고른다.","highlightBenefitJa":"振り分けの基になる複雑度判断を確認できる。","highlightBenefitKo":"라우팅에 사용한 복잡도 판단을 확인할 수 있다.","claimStatusJa":"固定版の README と実装を確認済み。このサイトでは独立した実行・性能検証をしていません。","claimStatusKo":"고정 버전 README와 구현을 확인했습니다. 이 사이트에서 독립적인 실행·성능 검증은 하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/BerriAI/litellm/blob/56116079c8022da0e8f7ff9ccb017ad5aca5aed2/litellm/router_strategy/complexity_router/jev_classifier.py#L70","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/BerriAI/litellm/blob/56116079c8022da0e8f7ff9ccb017ad5aca5aed2/README.md","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/BerriAI/litellm/blob/56116079c8022da0e8f7ff9ccb017ad5aca5aed2/tests/README.MD","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/BerriAI/litellm/blob/56116079c8022da0e8f7ff9ccb017ad5aca5aed2/LICENSE","note":"固定版本的实现与说明 / Pinned source and documentation"}],"pinned":false},{"id":"langchain-ai:langchainjs","name":"langchainjs","author":"langchain-ai","url":"https://github.com/langchain-ai/langchainjs","category":"SDK & Decision Frameworks","plainSummary":"LangChain.js 的可选 TypeSafeClassifier 集成，把状态和预设问题交给 Jev。","plainSummaryEn":"An optional LangChain.js TypeSafeClassifier integration for sending state and predefined questions to Jev.","jevDecisionPoint":"通过 invoke 调用 TypeSafe，解析 choice、noul、score 及其概率。","jevDecisionPointEn":"Uses invoke to call TypeSafe and parse choice, noul, score and probability fields.","highlightBenefit":"可把结构化判断接入 LangChain 流程，无需把它包装成聊天生成。","highlightBenefitEn":"Adds typed decisions to LangChain workflows without treating Jev as a chat generator.","tags":["multilanguage-sdk","classification-ranking"],"stars":18214,"forks":3380,"openIssues":568,"license":"MIT","language":"TypeScript","lastCommitAt":"2026-09-21T20:17:41Z","createdAt":"2023-02-06T22:50:27Z","metadataFetchedAt":"2026-09-21T21:58:35.937Z","avatarUrl":"https://avatars.githubusercontent.com/u/126733545?v=4","summarySource":"human-reviewed","claimStatus":"已核对固定版本 README 与调用源码。","claimStatusEn":"Reviewed the fixed-version README and call path.","plainSummaryJa":"状態と定義済み質問を Jev に送る LangChain.js の任意の TypeSafeClassifier 統合。","plainSummaryKo":"상태와 미리 정한 질문을 Jev에 보내는 선택적 LangChain.js TypeSafeClassifier 통합.","jevDecisionPointJa":"invoke から TypeSafe を呼び、choice・noul・score と確率を解析する。","jevDecisionPointKo":"invoke로 TypeSafe를 호출해 choice·noul·score와 확률을 파싱한다.","highlightBenefitJa":"Jev をチャット生成器に見立てず、LangChain の処理に型付き判断を加える。","highlightBenefitKo":"Jev를 채팅 생성기로 다루지 않고 LangChain 흐름에 구조화된 판단을 추가한다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/langchain-ai/langchainjs/blob/206d8b992bcf90ce7d46f2158f1ad85fc1d826c0/README.md","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/langchain-ai/langchainjs/blob/206d8b992bcf90ce7d46f2158f1ad85fc1d826c0/libs/providers/langchain-typesafe/README.md","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/langchain-ai/langchainjs/blob/206d8b992bcf90ce7d46f2158f1ad85fc1d826c0/libs/providers/langchain-typesafe/src/classifier.ts","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/langchain-ai/langchainjs/blob/206d8b992bcf90ce7d46f2158f1ad85fc1d826c0/libs/providers/langchain-typesafe/src/utils/client.ts","note":"固定版本说明与源码；本站未执行该项目。"}],"pinned":false},{"id":"kunchenguid:firstmate","name":"firstmate","author":"kunchenguid","url":"https://github.com/kunchenguid/firstmate","category":"Routing & Cost Optimization","plainSummary":"Firstmate 可选用 Jev 看任务简报并匹配派工规则，再由本地规则选择 Agent 配置。","plainSummaryEn":"Firstmate can optionally use Jev to match task briefs to dispatch rules before local policy chooses an Agent configuration.","jevDecisionPoint":"将任务简报与候选规则发送给 Jev，按匹配概率及本地条件解析执行配置。","jevDecisionPointEn":"Sends the task brief and candidate rules to Jev, then resolves execution profiles with confidence and local conditions.","highlightBenefit":"把语义匹配与最终派工规则分开。","highlightBenefitEn":"Separates semantic matching from the final dispatch policy.","tags":["llm-routing-cost","coding-agents"],"stars":6915,"forks":2109,"openIssues":1731,"license":"MIT","language":"Shell","lastCommitAt":"2026-09-21T19:08:50Z","createdAt":"2026-06-12T03:21:10Z","metadataFetchedAt":"2026-09-21T21:58:38.909Z","avatarUrl":"https://avatars.githubusercontent.com/u/3233006?v=4","summarySource":"human-reviewed","claimStatus":"这是 opt-in 派工辅助，不代表整个 Firstmate 工作流默认调用 Jev；任务简报会发送给 TypeSafe。","claimStatusEn":"This is opt-in dispatch assistance, not a claim that the whole Firstmate workflow uses Jev by default. Task briefs are sent to TypeSafe.","plainSummaryJa":"Firstmate は任意で Jev を使い、タスク概要と派工ルールを照合して Agent 設定を選ぶ。","plainSummaryKo":"Firstmate는 선택적으로 Jev로 작업 개요와 배분 규칙을 비교한 뒤 Agent 설정을 고른다.","jevDecisionPointJa":"概要と候補ルールを送り、確信度とローカル条件から実行設定を決める。","jevDecisionPointKo":"개요와 후보 규칙을 보내 신뢰도와 로컬 조건으로 실행 프로필을 결정한다.","highlightBenefitJa":"意味的な照合と最終的な派工ポリシーを分ける。","highlightBenefitKo":"의미 기반 매칭과 최종 배분 정책을 분리한다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"confirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/kunchenguid/firstmate/blob/4812db801628040b609dc25a2a8a91ed5efac662/bin/fm-dispatch-resolve.sh","note":"Fixed-commit source reviewed / 已核对固定版本源码"},{"url":"https://github.com/kunchenguid/firstmate/blob/4812db801628040b609dc25a2a8a91ed5efac662/README.md","note":"Fixed-commit source reviewed / 已核对固定版本源码"}],"pinned":false},{"id":"openchamber:openchamber","name":"openchamber","author":"openchamber","url":"https://github.com/openchamber/openchamber","category":"Routing & Cost Optimization","plainSummary":"OpenChamber 可选开启自动模型路由：Jev 看消息属于哪类任务，再使用该类绑定的模型和思考档位。","plainSummaryEn":"OpenChamber’s optional automatic model router uses Jev to classify a message before selecting a configured model and reasoning level.","jevDecisionPoint":"Choice 判断任务类别；另可用 Noul 判断自动批准的权限是否应留给用户确认。","jevDecisionPointEn":"Jev selects a task category; local category mappings determine the model configuration.","highlightBenefit":"把模型选择和权限提示接进现有会话界面，并记录失败回退原因。","highlightBenefitEn":"Keeps routing categories and their model assignments explicit.","tags":["llm-routing-cost","coding-agents"],"stars":10231,"forks":1126,"openIssues":638,"license":"MIT","language":"TypeScript","lastCommitAt":"2026-09-21T21:39:39Z","createdAt":"2025-09-11T10:34:59Z","metadataFetchedAt":"2026-09-21T21:58:41.916Z","avatarUrl":"https://avatars.githubusercontent.com/u/265872876?v=4","summarySource":"human-reviewed","claimStatus":"只核对可选路由模块；未验证具体模型组合的成本或质量改善。","claimStatusEn":"Only the optional routing module was inspected; cost or quality improvements for model combinations were not verified.","plainSummaryJa":"Jev でメッセージを分類し、設定済みモデルと思考レベルを選ぶ OpenChamber の任意ルーター。","plainSummaryKo":"Jev로 메시지를 분류해 설정된 모델과 추론 수준을 고르는 OpenChamber의 선택형 라우터.","jevDecisionPointJa":"Jev がタスク分類を選び、ローカルの対応表でモデル設定を決める。","jevDecisionPointKo":"Jev가 작업 범주를 선택하고 로컬 매핑이 모델 설정을 결정한다.","highlightBenefitJa":"分類とモデルの割り当てを明示できる。","highlightBenefitKo":"범주와 모델 배정을 명시적으로 관리한다.","claimStatusJa":"固定バージョンの文書とソースを確認。実行・性能・安全性は当サイトで独立検証していません。","claimStatusKo":"고정 버전 문서와 소스를 확인했습니다. 실행·성능·안전성은 이 사이트에서 독립적으로 검증하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/openchamber/openchamber/blob/614d7f76e581a132a86575c03d3fa9aad5e624b6/README.md","note":"README.md"},{"url":"https://github.com/openchamber/openchamber/blob/614d7f76e581a132a86575c03d3fa9aad5e624b6/packages/web/server/lib/routing/jev.js","note":"packages/web/server/lib/routing/jev.js"}],"pinned":false,"topics":["agentic-engineering","agents","ai","enginering","opencode"]},{"id":"ax-llm:ax","name":"ax","author":"ax-llm","url":"https://github.com/ax-llm/ax","category":"SDK & Integrations","plainSummary":"Ax 框架提供 TypeSafe 接口，可用布尔或有限类别签名调用 Jev，也可读原生答案。","plainSummaryEn":"Ax provides a TypeSafe integration for boolean or finite-class signatures and native Jev answers.","jevDecisionPoint":"将支持的签名映射到 Jev 问题，或直接发送原生 System One 请求。","jevDecisionPointEn":"Maps supported signatures to Jev questions or sends native System One requests.","highlightBenefit":"在 Ax 工作流中复用结构化签名与 Jev 概率结果。","highlightBenefitEn":"Reuses structured signatures and Jev probabilities within Ax workflows.","tags":["multilanguage-sdk","typed-decisions"],"stars":2940,"forks":195,"openIssues":8,"license":"Apache-2.0","lastCommitAt":"2026-09-18T07:45:03Z","createdAt":"2023-02-23T19:21:40Z","metadataFetchedAt":"2026-09-21T21:58:44.916Z","avatarUrl":"https://avatars.githubusercontent.com/u/172865539?v=4","summarySource":"human-reviewed","claimStatus":"已核对固定版本 README 与集成源码。","claimStatusEn":"README and integration source reviewed at a fixed commit; not independently run or benchmarked by this site.","plainSummaryJa":"Ax は真偽値や有限分類のシグネチャと、Jev のネイティブ回答に対応する TypeSafe 接続を提供する。","plainSummaryKo":"Ax는 불리언·유한 분류 시그니처와 Jev 원본 답변을 지원하는 TypeSafe 연결을 제공한다.","jevDecisionPointJa":"対応シグネチャを質問へ変換するか、System One リクエストを直接送る。","jevDecisionPointKo":"지원 시그니처를 질문으로 바꾸거나 System One 요청을 직접 보낸다.","highlightBenefitJa":"Ax の処理内でシグネチャと Jev の確率結果を利用できる。","highlightBenefitKo":"Ax 흐름에서 구조화된 시그니처와 Jev 확률을 활용할 수 있다.","claimStatusJa":"固定版の README と実装を確認済み。このサイトでは独立した実行・性能検証をしていません。","claimStatusKo":"고정 버전 README와 구현을 확인했습니다. 이 사이트에서 독립적인 실행·성능 검증은 하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/ax-llm/ax/blob/5c43344f9ef3016db576fa2c3b59d48ef21b4d71/src/ax/ai/typesafe/client.ts#L1","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/ax-llm/ax/blob/5c43344f9ef3016db576fa2c3b59d48ef21b4d71/README.md","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/ax-llm/ax/blob/5c43344f9ef3016db576fa2c3b59d48ef21b4d71/src/ax/README.md","note":"固定版本的实现与说明 / Pinned source and documentation"}],"pinned":false},{"id":"agentgateway:agentgateway","name":"agentgateway","author":"agentgateway","url":"https://github.com/agentgateway/agentgateway","category":"Security & Guardrails","plainSummary":"Agentgateway 仓库里的 Jev 护栏示例，通过 webhook 检查模型请求和回复。","plainSummaryEn":"A Jev guardrail example in Agentgateway using a webhook to inspect model requests and responses.","jevDecisionPoint":"Jev 对越狱、有害内容和秘密泄露评分，示例按阈值或评估错误拒绝请求。","jevDecisionPointEn":"Jev scores jailbreaks, harmful content and secret disclosure; thresholds or evaluation errors reject requests.","highlightBenefit":"展示网关接入方式；模型评分不能保证阻挡全部攻击。","highlightBenefitEn":"Demonstrates gateway integration, not guaranteed prevention of all attacks.","tags":["security-guardrails","multilanguage-sdk"],"stars":4965,"forks":865,"openIssues":285,"license":"Apache-2.0","language":"Rust","lastCommitAt":"2026-09-21T21:28:42Z","createdAt":"2025-03-18T20:55:22Z","metadataFetchedAt":"2026-09-21T21:58:47.923Z","avatarUrl":"https://avatars.githubusercontent.com/u/208695296?v=4","summarySource":"human-reviewed","claimStatus":"已核对固定版本 README 与调用源码。","claimStatusEn":"Reviewed the fixed-version README and call path.","plainSummaryJa":"Agentgateway に含まれる Jev ガードレール例。webhook で要求と応答を点検する。","plainSummaryKo":"webhook으로 모델 요청·응답을 점검하는 Agentgateway의 Jev 가드레일 예제.","jevDecisionPointJa":"脱獄・有害内容・秘密漏えいを採点し、閾値や評価エラーで要求を拒否する。","jevDecisionPointKo":"탈옥·유해 내용·비밀 유출을 채점하고 임곗값이나 평가 오류에 따라 요청을 거부한다.","highlightBenefitJa":"ゲートウェイ統合の例で、すべての攻撃の防止を保証しない。","highlightBenefitKo":"게이트웨이 통합 예제이며 모든 공격 차단을 보장하지 않는다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/agentgateway/agentgateway/blob/6b0270efd25b5255932943e48b5ca47583d3ad28/README.md","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/agentgateway/agentgateway/blob/6b0270efd25b5255932943e48b5ca47583d3ad28/examples/llm-guardrail-jev/README.md","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/agentgateway/agentgateway/blob/6b0270efd25b5255932943e48b5ca47583d3ad28/examples/llm-guardrail-jev/guardrail.ts","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/agentgateway/agentgateway/blob/6b0270efd25b5255932943e48b5ca47583d3ad28/examples/llm-guardrail-jev/config.yaml","note":"固定版本说明与源码；本站未执行该项目。"}],"pinned":false,"topics":["agents","ai","ai-gateway","api-gateway","gateway-api","kubernetes","mcp","mcp-gateway","reverse-proxy","rust","service-mesh"]},{"id":"bennykok:omg.dev","name":"omg.dev","author":"BennyKok","url":"https://github.com/BennyKok/omg.dev","category":"Browser & OS Action","plainSummary":"omg.dev 的移动端测试脚本可让 Jev 读取可访问性树并选择下一步交互。","plainSummaryEn":"An omg.dev mobile testing script can use Jev to read the accessibility tree and choose the next interaction.","jevDecisionPoint":"判断要操作的控件、步骤是否完成或是否受阻，再由测试执行器操作界面。","jevDecisionPointEn":"Chooses controls and checks completion or blockage before the test runner operates the UI.","highlightBenefit":"给移动端测试加入基于当前界面状态的选择。","highlightBenefitEn":"Adds choices grounded in current UI state to mobile testing.","tags":["desktop-os","evaluation-benchmarks"],"stars":533,"forks":40,"openIssues":6,"license":"MIT","language":"TypeScript","lastCommitAt":"2026-09-21T18:00:39Z","createdAt":"2026-06-17T07:33:04Z","metadataFetchedAt":"2026-09-21T21:58:50.950Z","avatarUrl":"https://avatars.githubusercontent.com/u/18395202?v=4","summarySource":"human-reviewed","claimStatus":"只核实移动测试脚本的可选接入，不能描述成 omg.dev 全产品都由 Jev 控制；本站未执行测试。","claimStatusEn":"Only the optional mobile-test integration was verified, not that the entire omg.dev product is controlled by Jev. This site has not executed the tests.","plainSummaryJa":"omg.dev のモバイルテスト用スクリプトが、アクセシビリティツリーから次の操作を Jev に選ばせる。","plainSummaryKo":"omg.dev의 모바일 테스트 스크립트가 접근성 트리에서 다음 동작을 Jev로 선택한다.","jevDecisionPointJa":"対象や完了、行き詰まりを判断し、テスト実行器が画面を操作する。","jevDecisionPointKo":"대상, 완료 여부, 진행 불가 상태를 판단하면 테스트 실행기가 화면을 조작한다.","highlightBenefitJa":"現在の画面状態に基づく選択をモバイルテストへ加える。","highlightBenefitKo":"현재 화면 상태에 기반한 선택을 모바일 테스트에 추가한다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"confirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/BennyKok/omg.dev/blob/a00f56684a569ee417be787c148eaa9874946f2b/mobile/scripts/jev.ts","note":"Fixed-commit source reviewed / 已核对固定版本源码"},{"url":"https://github.com/BennyKok/omg.dev/blob/a00f56684a569ee417be787c148eaa9874946f2b/README.md","note":"Fixed-commit source reviewed / 已核对固定版本源码"}],"pinned":false},{"id":"monotykamary:pi-fabric","name":"pi-fabric","author":"monotykamary","url":"https://github.com/monotykamary/pi-fabric","category":"SDK & Decision Frameworks","plainSummary":"给 Pi 的工具运行时加上可编程的 Jev 决策循环。先写好观察、判断和执行步骤，再让它按预算在前台或后台运行。","plainSummaryEn":"Pi’s programmable runtime includes an optional Jev loop for observing state, making decisions and running bounded actions.","jevDecisionPoint":"对程序提交的状态做单选、是非判断或评分；循环和动作执行由本地程序控制。","jevDecisionPointEn":"Programmed questions feed Jev judgments into local action logic under runtime budgets.","highlightBenefit":"适合把反复发生的小判断写成可复用流程；Jev 是需要配置的可选能力。","highlightBenefitEn":"Keeps observation, judgment and execution steps in a reusable program.","tags":["coding-agents","mcp-integrations","typed-decisions"],"stars":244,"forks":25,"openIssues":4,"license":"MIT","language":"TypeScript","lastCommitAt":"2026-09-21T17:17:52Z","createdAt":"2026-07-13T14:35:18Z","metadataFetchedAt":"2026-09-21T21:58:53.950Z","avatarUrl":"https://avatars.githubusercontent.com/u/1130103?v=4","summarySource":"human-reviewed","claimStatus":"这是 pi-fabric 中的可选 Jev 能力，不将仓库其他功能或基准成绩归功于 Jev。","claimStatusEn":"This covers the optional Jev capability in pi-fabric; other features and benchmark results are not attributed to Jev. Not run here.","plainSummaryJa":"Pi のプログラム可能な runtime に、観測・Jev 判断・制限付き実行のループを追加する。","plainSummaryKo":"Pi의 프로그래밍 가능한 runtime에 관측·Jev 판단·제한된 실행 루프를 추가한다.","jevDecisionPointJa":"記述した質問の回答を、予算制限付きのローカル実行ロジックへ渡す。","jevDecisionPointKo":"작성한 질문의 답을 예산 제한이 있는 로컬 동작 로직에 전달한다.","highlightBenefitJa":"観測・判断・実行を再利用可能なプログラムとして定義できる。","highlightBenefitKo":"관측·판단·실행을 재사용 가능한 프로그램으로 정의한다.","claimStatusJa":"固定バージョンの文書とソースを確認。実行・性能・安全性は当サイトで独立検証していません。","claimStatusKo":"고정 버전 문서와 소스를 확인했습니다. 실행·성능·안전성은 이 사이트에서 독립적으로 검증하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/monotykamary/pi-fabric/blob/2e5fcdd94cc364445b115756edefcaab24f68928/README.md","note":"README.md"},{"url":"https://github.com/monotykamary/pi-fabric/blob/2e5fcdd94cc364445b115756edefcaab24f68928/src/jev/client.ts","note":"src/jev/client.ts"}],"pinned":false,"topics":["agents","coding-agent","mcp","multi-agent","orchestration","pi","pi-extension","pi-package","quickjs","rlm","typescript"]},{"id":"latitude-dev:latitude-llm","name":"latitude-llm","author":"latitude-dev","url":"https://github.com/latitude-dev/latitude-llm","category":"Evaluation & Observability","plainSummary":"Latitude 的可选 Jev 预分类器为对话检查打分，并记录检查选择的依据。","plainSummaryEn":"Latitude includes an optional Jev preclassifier for conversation checks and their selection records.","jevDecisionPoint":"判断各项检查的适用性，满足阈值与限流条件时补充检查任务。","jevDecisionPointEn":"Judges which checks apply and can add checks when thresholds and rate limits permit.","highlightBenefit":"记录模型、阈值、调用时间与选择原因，便于对照原流程。","highlightBenefitEn":"Records models, thresholds, latency and selection reasons alongside the baseline.","tags":["evaluation-benchmarks","voice-conversation"],"stars":4665,"forks":392,"openIssues":89,"license":"MIT","lastCommitAt":"2026-09-21T17:09:59Z","createdAt":"2024-06-28T11:16:02Z","metadataFetchedAt":"2026-09-21T21:58:56.923Z","avatarUrl":"https://avatars.githubusercontent.com/u/98949449?v=4","summarySource":"human-reviewed","claimStatus":"已核对固定版本 README 与集成源码。","claimStatusEn":"README and integration source reviewed at a fixed commit; not independently run or benchmarked by this site.","plainSummaryJa":"Latitude の任意の Jev 事前分類器が、会話検査の判断と選択理由を記録する。","plainSummaryKo":"Latitude의 선택형 Jev 사전 분류기가 대화 검사 판단과 선택 근거를 기록한다.","jevDecisionPointJa":"検査の必要性を判断し、しきい値と呼び出し制限を満たす場合に検査を追加する。","jevDecisionPointKo":"검사의 필요성을 판단하고 임계값과 호출 제한을 만족하면 검사를 추가한다.","highlightBenefitJa":"モデル、しきい値、時間、選択理由を元の処理と対照できる。","highlightBenefitKo":"모델, 임계값, 지연과 선택 이유를 기존 처리와 비교할 수 있다.","claimStatusJa":"固定版の README と実装を確認済み。このサイトでは独立した実行・性能検証をしていません。","claimStatusKo":"고정 버전 README와 구현을 확인했습니다. 이 사이트에서 독립적인 실행·성능 검증은 하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/latitude-dev/latitude-llm/blob/6b2484c1d74973877fc7d353b8867c7496568d78/packages/domain/flaggers/src/use-cases/run-jev-preclassifier.ts#L1","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/latitude-dev/latitude-llm/blob/6b2484c1d74973877fc7d353b8867c7496568d78/README.md","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/latitude-dev/latitude-llm/blob/6b2484c1d74973877fc7d353b8867c7496568d78/fern/README.md","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/latitude-dev/latitude-llm/blob/6b2484c1d74973877fc7d353b8867c7496568d78/packages/platform/ai-jev/src/jev-shadow-decision-provider.ts","note":"固定版本的实现与说明 / Pinned source and documentation"}],"pinned":false},{"id":"cognesy:instructor-php","name":"instructor-php","author":"cognesy","url":"https://github.com/cognesy/instructor-php","category":"SDK & Decision Frameworks","plainSummary":"Instructor PHP 的 Polyglot 模块内置 TypeSafe Decision 驱动。","plainSummaryEn":"A TypeSafe Decision driver within Instructor PHP’s Polyglot module.","jevDecisionPoint":"将业务状态和类型化问题转换为 Jev 请求，再映射为 PHP 决策响应。","jevDecisionPointEn":"Converts application state and typed questions into Jev requests and maps responses to PHP decision objects.","highlightBenefit":"PHP 应用可通过统一 Decision 接口使用 Jev，而不自行拼装 HTTP 数据。","highlightBenefitEn":"Lets PHP applications use Jev through a shared Decision interface.","tags":["multilanguage-sdk","typed-decisions"],"stars":327,"forks":27,"openIssues":2,"license":"MIT","language":"PHP","lastCommitAt":"2026-09-17T17:04:53Z","createdAt":"2024-02-26T10:47:00Z","metadataFetchedAt":"2026-09-21T21:58:59.953Z","avatarUrl":"https://avatars.githubusercontent.com/u/128147984?v=4","summarySource":"human-reviewed","claimStatus":"已核对固定版本 README 与调用源码。","claimStatusEn":"Reviewed the fixed-version README and call path.","plainSummaryJa":"Instructor PHP の Polyglot モジュールにある TypeSafe Decision ドライバー。","plainSummaryKo":"Instructor PHP의 Polyglot 모듈에 포함된 TypeSafe Decision 드라이버.","jevDecisionPointJa":"状態と型付き質問を Jev 要求へ変換し、応答を PHP の判断オブジェクトへ写像する。","jevDecisionPointKo":"상태와 구조화된 질문을 Jev 요청으로 변환하고 응답을 PHP 판단 객체로 매핑한다.","highlightBenefitJa":"共通の Decision インターフェースから PHP アプリで Jev を利用できる。","highlightBenefitKo":"PHP 앱에서 공통 Decision 인터페이스로 Jev를 사용할 수 있다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/cognesy/instructor-php/blob/bb1160ce2360dcf1434788c46ee97809fe782ea3/README.md","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/cognesy/instructor-php/blob/bb1160ce2360dcf1434788c46ee97809fe782ea3/packages/polyglot/src/Decision/Drivers/TypeSafe/TypesafeDriver.php","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/cognesy/instructor-php/blob/bb1160ce2360dcf1434788c46ee97809fe782ea3/packages/polyglot/src/Decision/Drivers/TypeSafe/TypesafeRequestAdapter.php","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/cognesy/instructor-php/blob/bb1160ce2360dcf1434788c46ee97809fe782ea3/packages/polyglot/src/Decision/Drivers/TypeSafe/TypesafeResponseAdapter.php","note":"固定版本说明与源码；本站未执行该项目。"}],"pinned":false,"topics":["agents","ai","ai-agents","genai","llm","php"]},{"id":"bastani-inc:atomic","name":"atomic","author":"bastani-inc","url":"https://github.com/bastani-inc/atomic","category":"Routing & Cost Optimization","plainSummary":"Atomic 编程 Agent 的可选 Jev 决策后端，为路由等流程提供受限的结构化选择。","plainSummaryEn":"An optional Jev decision backend in the Atomic coding Agent for bounded structured choices such as routing.","jevDecisionPoint":"将预定义的问题交给 Jev，解码答案后由调用方应用；常规模型仍负责生成代码。","jevDecisionPointEn":"Sends predefined questions to Jev and decodes answers for callers; regular models still generate code.","highlightBenefit":"让结构化决策与文本生成使用各自的接口。","highlightBenefitEn":"Separates structured decisions from text-generation interfaces.","tags":["coding-agents","llm-routing-cost"],"stars":809,"forks":112,"openIssues":38,"license":null,"language":"TypeScript","lastCommitAt":"2026-09-21T17:56:15Z","createdAt":"2025-10-23T04:24:21Z","metadataFetchedAt":"2026-09-21T21:59:02.972Z","avatarUrl":"https://avatars.githubusercontent.com/u/252341057?v=4","summarySource":"human-reviewed","claimStatus":"核实的是可选 structured-output provider，不表示 Atomic 默认或全部功能依赖 Jev。 LICENSE 另含商业规模达到阈值后的界面署名条件，非标准 MIT，请阅读原文。","claimStatusEn":"The verified scope is an optional structured-output provider, not that Atomic uses Jev by default or for every feature. LICENSE includes an additional UI-attribution condition above stated business-size thresholds. It is not standard MIT; read the license text.","plainSummaryJa":"Atomic のコーディング Agent にある任意の Jev バックエンドで、振り分けなどの構造化選択を行う。","plainSummaryKo":"Atomic 코딩 Agent의 선택적 Jev 백엔드로 라우팅 등 제한된 구조화 선택을 수행한다.","jevDecisionPointJa":"事前定義の質問を送り回答を呼び出し元で使い、コード生成は通常のモデルが担当する。","jevDecisionPointKo":"미리 정의한 질문을 보내 답을 호출자가 사용하며 코드 생성은 일반 모델이 맡는다.","highlightBenefitJa":"構造化判断とテキスト生成のインターフェースを分ける。","highlightBenefitKo":"구조화된 판단과 텍스트 생성 인터페이스를 분리한다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"custom","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/bastani-inc/atomic/blob/33ca4ccb5f39488aeba67c0212a52a31cd14059d/packages/coding-agent/src/core/structured-output/jev.ts","note":"Fixed-commit source reviewed / 已核对固定版本源码"},{"url":"https://github.com/bastani-inc/atomic/blob/33ca4ccb5f39488aeba67c0212a52a31cd14059d/README.md","note":"Fixed-commit source reviewed / 已核对固定版本源码"},{"url":"https://github.com/bastani-inc/atomic/blob/33ca4ccb5f39488aeba67c0212a52a31cd14059d/LICENSE","note":"Fixed-commit source reviewed / 已核对固定版本源码"}],"pinned":false},{"id":"wrongstack:wrongstack","name":"WrongStack","author":"WrongStack","url":"https://github.com/WrongStack/WrongStack","category":"Routing & Cost Optimization","plainSummary":"给 WrongStack 编程 Agent 增加一个可选分派助手。遇到多个相近的专长 Agent 时，用 Jev 判断谁更适合当前任务。","plainSummaryEn":"An optional Jev dispatch classifier for choosing among WrongStack specialist Agents.","jevDecisionPoint":"从候选角色中选一个，同时判断有没有任何候选真正适合；不合适时交回原有策略。","jevDecisionPointEn":"Jev evaluates the task against eligible specialists; local dispatch rules use the result.","highlightBenefit":"分派结果带有可检查的概率，也能明确表示没有合适人选。","highlightBenefitEn":"Makes specialist selection a configurable step in an existing coding Agent.","tags":["coding-agents","llm-routing-cost"],"stars":331,"forks":37,"openIssues":13,"license":"MIT","language":"TypeScript","lastCommitAt":"2026-09-21T15:09:26Z","createdAt":"2026-05-12T23:57:23Z","metadataFetchedAt":"2026-09-21T21:59:05.953Z","avatarUrl":"https://avatars.githubusercontent.com/u/284111525?v=4","summarySource":"human-reviewed","claimStatus":"可选分派模块，不能据此声称所有 WrongStack 任务采用 Jev 或保证正确派发。","claimStatusEn":"An optional dispatch module, not evidence that all WrongStack tasks use Jev or that dispatch is always correct.","plainSummaryJa":"WrongStack の専門 Agent を選ぶための、任意の Jev 分類器。","plainSummaryKo":"WrongStack의 전문 Agent를 선택하는 선택형 Jev 분류기.","jevDecisionPointJa":"適格な専門 Agent とタスクを照合し、ローカル配分規則が結果を使う。","jevDecisionPointKo":"적격 전문 Agent와 작업을 비교하고 로컬 배분 규칙이 결과를 사용한다.","highlightBenefitJa":"既存コーディング Agent に設定可能な専門家選択を追加できる。","highlightBenefitKo":"기존 코딩 Agent에 설정 가능한 전문가 선택 단계를 추가한다.","claimStatusJa":"固定バージョンの文書とソースを確認。実行・性能・安全性は当サイトで独立検証していません。","claimStatusKo":"고정 버전 문서와 소스를 확인했습니다. 실행·성능·안전성은 이 사이트에서 독립적으로 검증하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/WrongStack/WrongStack/blob/4cf97c0aa4f855751949ee8e99c719d06f4c26e0/README.md","note":"README.md"},{"url":"https://github.com/WrongStack/WrongStack/blob/4cf97c0aa4f855751949ee8e99c719d06f4c26e0/packages/core/src/types/config/typesafe.ts","note":"packages/core/src/types/config/typesafe.ts"},{"url":"https://github.com/WrongStack/WrongStack/blob/4cf97c0aa4f855751949ee8e99c719d06f4c26e0/packages/core/src/coordination/typesafe-dispatch-classifier.ts","note":"packages/core/src/coordination/typesafe-dispatch-classifier.ts"}],"pinned":false,"topics":["agentic-ai","ai-agents","ai-tools","code","code-generation","coding"]},{"id":"celestoai:celesto","name":"celesto","author":"CelestoAI","url":"https://github.com/CelestoAI/celesto","category":"Codebase & Graph Pathfinding","plainSummary":"Celesto 的 PR 审查示例在沙盒中准备检查，再比较普通模型与 Jev 对候选问题的判断。","plainSummaryEn":"A Celesto PR-review example prepares sandbox checks and compares a general model with Jev on candidate findings.","jevDecisionPoint":"检查疑似问题是否由当前变更引入、是否有证据、是否值得修复。","jevDecisionPointEn":"Judges whether a finding was introduced by the change, is supported and merits a fix.","highlightBenefit":"将执行记录与评估结果放在同一审查界面中。","highlightBenefitEn":"Shows execution records and reviewer judgments together.","tags":["coding-agents","evaluation-benchmarks"],"stars":957,"forks":75,"openIssues":18,"license":"Apache-2.0","lastCommitAt":"2026-09-21T06:42:32Z","createdAt":"2026-02-15T00:31:08Z","metadataFetchedAt":"2026-09-21T21:59:08.934Z","avatarUrl":"https://avatars.githubusercontent.com/u/187575487?v=4","summarySource":"human-reviewed","claimStatus":"已核对固定版本 README 与集成源码。","claimStatusEn":"README and integration source reviewed at a fixed commit; not independently run or benchmarked by this site.","plainSummaryJa":"Celesto の PR レビュー例がサンドボックスで検査を準備し、通常モデルと Jev の指摘評価を比較する。","plainSummaryKo":"Celesto PR 검토 예제가 샌드박스 검사를 준비하고 일반 모델과 Jev의 후보 문제 판단을 비교한다.","jevDecisionPointJa":"問題が今回の変更によるものか、根拠があるか、修正対象かを判断する。","jevDecisionPointKo":"문제가 이번 변경에서 생겼는지, 근거가 있는지, 수정할 가치가 있는지 판단한다.","highlightBenefitJa":"実行記録とレビュー判断を同じ画面で確認できる。","highlightBenefitKo":"실행 기록과 검토 판단을 함께 보여 준다.","claimStatusJa":"固定版の README と実装を確認済み。このサイトでは独立した実行・性能検証をしていません。","claimStatusKo":"고정 버전 README와 구현을 확인했습니다. 이 사이트에서 독립적인 실행·성능 검증은 하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/CelestoAI/celesto/blob/fff7bb567752baee4ce195fcb78b5ee5e4da889f/examples/pr-review-jev/models.py#L96","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/CelestoAI/celesto/blob/fff7bb567752baee4ce195fcb78b5ee5e4da889f/examples/pr-review-jev/README.md","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/CelestoAI/celesto/blob/fff7bb567752baee4ce195fcb78b5ee5e4da889f/README.md","note":"固定版本的实现与说明 / Pinned source and documentation"}],"pinned":false},{"id":"agentjido:req_llm","name":"req_llm","author":"agentjido","url":"https://github.com/agentjido/req_llm","category":"SDK & Decision Frameworks","plainSummary":"ReqLLM 的 TypeSafe provider，让 Elixir 应用通过 evaluate 接口调用 Jev。","plainSummaryEn":"A TypeSafe provider for calling Jev through ReqLLM’s evaluate interface in Elixir.","jevDecisionPoint":"提交状态与判断题，将答案放入统一响应对象，并保留原始 provider 数据。","jevDecisionPointEn":"Sends state and questions, normalizes answers and retains the raw provider response.","highlightBenefit":"把决策 API 与聊天生成分开，便于在 Elixir 程序中处理结果。","highlightBenefitEn":"Keeps decision evaluation separate from chat generation in Elixir applications.","tags":["multilanguage-sdk","typed-decisions"],"stars":581,"forks":198,"openIssues":3,"license":"Apache-2.0","language":"Elixir","lastCommitAt":"2026-09-21T20:56:55Z","createdAt":"2025-09-05T20:23:26Z","metadataFetchedAt":"2026-09-21T21:59:11.953Z","avatarUrl":"https://avatars.githubusercontent.com/u/191822461?v=4","summarySource":"human-reviewed","claimStatus":"已核对固定版本 README 与调用源码。","claimStatusEn":"Reviewed the fixed-version README and call path.","plainSummaryJa":"Elixir の ReqLLM evaluate インターフェースから Jev を呼ぶ TypeSafe provider。","plainSummaryKo":"Elixir ReqLLM의 evaluate 인터페이스로 Jev를 호출하는 TypeSafe provider.","jevDecisionPointJa":"状態と質問を送り、回答を正規化しつつ provider の生データも保持する。","jevDecisionPointKo":"상태와 질문을 전송하고 응답을 정규화하면서 provider 원본도 보존한다.","highlightBenefitJa":"判断評価をチャット生成から分離して扱える。","highlightBenefitKo":"판단 평가를 채팅 생성과 분리해 처리할 수 있다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/agentjido/req_llm/blob/9cb0ee7a0fea5f3520fc953911d352c93193615e/README.md","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/agentjido/req_llm/blob/9cb0ee7a0fea5f3520fc953911d352c93193615e/lib/req_llm/providers/typesafe.ex","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/agentjido/req_llm/blob/9cb0ee7a0fea5f3520fc953911d352c93193615e/lib/req_llm/evaluation.ex","note":"固定版本说明与源码；本站未执行该项目。"}],"pinned":false,"topics":["ai","ai-providers","elixir","elixir-package","llm","req"]},{"id":"vellum-ai:vellum-assistant","name":"vellum-assistant","author":"vellum-ai","url":"https://github.com/vellum-ai/vellum-assistant","category":"MCP & Integrations","plainSummary":"Vellum Assistant 中的可选 Jev provider，可把会话状态与明确的问题交给 TypeSafe。","plainSummaryEn":"An optional Jev provider in Vellum Assistant sends conversation state and explicit questions to TypeSafe.","jevDecisionPoint":"将状态和问题包发送到 System One，再把结构化答案返回给 Assistant。","jevDecisionPointEn":"Submits state and question bundles to System One and returns structured answers to the Assistant.","highlightBenefit":"在已有助手中接入选择、概率和评分判断。","highlightBenefitEn":"Adds choices, probabilities, and scoring to an existing assistant.","tags":["mcp-integrations","typed-decisions"],"stars":1295,"forks":189,"openIssues":170,"license":"MIT","language":"TypeScript","lastCommitAt":"2026-09-21T21:51:03Z","createdAt":"2026-02-07T21:19:17Z","metadataFetchedAt":"2026-09-21T21:59:14.962Z","avatarUrl":"https://avatars.githubusercontent.com/u/118817114?v=4","summarySource":"human-reviewed","claimStatus":"只核实 Jev provider 的接入范围，不代表整个助手的记忆、自动化或行为由 Jev 驱动。","claimStatusEn":"Only the Jev provider integration was verified, not that Jev drives the assistant’s memory, automation, or overall behavior.","plainSummaryJa":"Vellum Assistant の任意の Jev provider が、会話状態と明示的な質問を TypeSafe に送る。","plainSummaryKo":"Vellum Assistant의 선택적 Jev provider가 대화 상태와 명시적 질문을 TypeSafe에 보낸다.","jevDecisionPointJa":"状態と質問を System One に送り、構造化された回答を Assistant に返す。","jevDecisionPointKo":"상태와 질문 묶음을 System One에 보내 구조화된 답을 Assistant로 반환한다.","highlightBenefitJa":"既存のアシスタントに選択、確率、採点を追加できる。","highlightBenefitKo":"기존 어시스턴트에 선택, 확률, 평가 기능을 추가할 수 있다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"confirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/vellum-ai/vellum-assistant/blob/ee5ba342719e72b67698c2e1a4a78837321d0b1d/assistant/src/providers/jev/client.ts","note":"Fixed-commit source reviewed / 已核对固定版本源码"},{"url":"https://github.com/vellum-ai/vellum-assistant/blob/ee5ba342719e72b67698c2e1a4a78837321d0b1d/README.md","note":"Fixed-commit source reviewed / 已核对固定版本源码"}],"pinned":false,"topics":["agent","agentic-ai","ai-assistant","anthropic","autonomous-agents","claude","claude-code","ios","macos","memory","open-source","openai","openclaw","proactive-ai","slack","telegram-bot","vellum-ai"]},{"id":"notque:vexjoy-agent","name":"vexjoy-agent","author":"notque","url":"https://github.com/notque/vexjoy-agent","category":"Routing & Cost Optimization","plainSummary":"给 VexJoy 的任务分派增加一条 Jev 路线。输入需求后，判断该选哪位专长 Agent、哪项技能和哪条工作流。","plainSummaryEn":"An optional Jev routing path that matches VexJoy requests to specialist Agents, skills and workflows.","jevDecisionPoint":"先判断是否需要路由，再从实际清单中选择候选；非法选择或调用失败交回原有流程。","jevDecisionPointEn":"After deterministic routing guards, Jev judges the remaining candidates and required workflow components.","highlightBenefit":"提供可对照的路由实验入口；项目文档明确说当前评测不足以把它提升为默认路线。","highlightBenefitEn":"Keeps repeatable routing rules and model-selected candidates in separate stages.","tags":["llm-routing-cost","mcp-integrations"],"stars":422,"forks":45,"openIssues":0,"license":"MIT","language":"Python","lastCommitAt":"2026-09-21T21:01:02Z","createdAt":"2026-03-18T01:39:09Z","metadataFetchedAt":"2026-09-21T21:59:17.949Z","avatarUrl":"https://avatars.githubusercontent.com/u/220916?v=4","summarySource":"human-reviewed","claimStatus":"存在强制路由与不可用回退路径；不将所有分派归因于 Jev。","claimStatusEn":"Forced routing and unavailable-provider paths exist; not every dispatch is a Jev decision. Not run here.","plainSummaryJa":"VexJoy の依頼を専門 Agent・skill・workflow に対応付ける任意の Jev 経路。","plainSummaryKo":"VexJoy 요청을 전문 Agent·skill·workflow에 연결하는 선택형 Jev 경로.","jevDecisionPointJa":"決定的な振り分け規則の後で、残りの候補と必要な構成要素を判定する。","jevDecisionPointKo":"결정적 라우팅 규칙을 거친 뒤 남은 후보와 필요한 구성 요소를 판단한다.","highlightBenefitJa":"固定規則とモデルによる候補選択を別の段階に分ける。","highlightBenefitKo":"고정 규칙과 모델의 후보 선택을 별도 단계로 구분한다.","claimStatusJa":"固定バージョンの文書とソースを確認。実行・性能・安全性は当サイトで独立検証していません。","claimStatusKo":"고정 버전 문서와 소스를 확인했습니다. 실행·성능·안전성은 이 사이트에서 독립적으로 검증하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/notque/vexjoy-agent/blob/ab51ee7da567e84f72f76d73b6e7b83cec5ef060/README.md","note":"README.md"},{"url":"https://github.com/notque/vexjoy-agent/blob/ab51ee7da567e84f72f76d73b6e7b83cec5ef060/scripts/jev-route.py","note":"scripts/jev-route.py"}],"pinned":false,"topics":["ai","ai-agents","ai-pipelines","ai-skills","ai-workflows","claude","claude-ai","claude-code","claude-code-skill","claude-skills","codex-skills","developer-tools","factory-ai"]},{"id":"uezo:aiavatarkit","name":"aiavatarkit","author":"uezo","url":"https://github.com/uezo/aiavatarkit","category":"Voice & Conversation","plainSummary":"AIAvatarKit 的可选 Jev 组件根据转写内容判断用户是否结束发言。","plainSummaryEn":"An optional AIAvatarKit component uses Jev to judge turn endings from speech transcripts.","jevDecisionPoint":"评估当前话语是否完整、是否准备继续说，并影响轮次结束门槛。","jevDecisionPointEn":"Assesses utterance completeness and whether the user is likely to continue speaking.","highlightBenefit":"给仅靠静音时长的轮次判断增加语义信号。","highlightBenefitEn":"Adds a semantic signal to silence-based turn detection.","tags":["voice-conversation","typed-decisions"],"stars":678,"forks":66,"openIssues":7,"license":"Apache-2.0","lastCommitAt":"2026-09-17T15:48:49Z","createdAt":"2023-05-27T14:22:56Z","metadataFetchedAt":"2026-09-21T21:59:20.940Z","avatarUrl":"https://avatars.githubusercontent.com/u/19625242?v=4","summarySource":"human-reviewed","claimStatus":"已核对固定版本 README 与集成源码。","claimStatusEn":"README and integration source reviewed at a fixed commit; not independently run or benchmarked by this site.","plainSummaryJa":"AIAvatarKit の任意の部品が、音声の書き起こしから発話終了を Jev に判断させる。","plainSummaryKo":"AIAvatarKit의 선택형 구성요소가 음성 전사로 발화 종료를 Jev에 판단시킨다.","jevDecisionPointJa":"発話の完結性と、利用者が話し続けそうかを評価する。","jevDecisionPointKo":"발화가 완결됐는지, 사용자가 계속 말할 가능성이 있는지 평가한다.","highlightBenefitJa":"無音時間による判定に意味上の情報を加える。","highlightBenefitKo":"무음 시간 기반 판정에 의미 정보를 더한다.","claimStatusJa":"固定版の README と実装を確認済み。このサイトでは独立した実行・性能検証をしていません。","claimStatusKo":"고정 버전 README와 구현을 확인했습니다. 이 사이트에서 독립적인 실행·성능 검증은 하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/uezo/aiavatarkit/blob/38b617b8b9269939734e70ef503d7ea6976acdbd/aiavatar/sts/vad/turn_end_gates/jev.py#L172","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/uezo/aiavatarkit/blob/38b617b8b9269939734e70ef503d7ea6976acdbd/README.md","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/uezo/aiavatarkit/blob/38b617b8b9269939734e70ef503d7ea6976acdbd/examples/nod/README.md","note":"固定版本的实现与说明 / Pinned source and documentation"}],"pinned":false},{"id":"ldbumble:taskuary","name":"taskuary","author":"ldbumble","url":"https://github.com/ldbumble/taskuary","category":"Evaluation & Observability","plainSummary":"Taskuary 的可选 Jev 判断模块，对任务运行状态检查用户定义的条件。","plainSummaryEn":"An optional Jev judgment module in Taskuary for checking user-defined conditions on task state.","jevDecisionPoint":"把条件转成 yes/no 概率问题，按本地阈值返回布尔结果和原始概率。","jevDecisionPointEn":"Turns conditions into yes/no probability questions and returns local-threshold verdicts plus probabilities.","highlightBenefit":"可为任务结果增加结构化检查；不是整个消息系统都由 Jev 控制。","highlightBenefitEn":"Adds structured checks to task outcomes without making Jev the controller of the whole messaging system.","tags":["evaluation-benchmarks","typed-decisions"],"stars":116,"forks":21,"openIssues":9,"license":"MIT","language":"Python","lastCommitAt":"2026-09-20T20:37:08Z","createdAt":"2026-08-17T12:53:50Z","metadataFetchedAt":"2026-09-21T21:59:23.956Z","avatarUrl":"https://avatars.githubusercontent.com/u/116741397?v=4","summarySource":"human-reviewed","claimStatus":"已核对固定版本 README 与调用源码。","claimStatusEn":"Reviewed the fixed-version README and call path.","plainSummaryJa":"Taskuary でタスク状態がユーザー定義条件を満たすか点検する任意の Jev モジュール。","plainSummaryKo":"Taskuary에서 작업 상태가 사용자 정의 조건을 충족하는지 확인하는 선택적 Jev 모듈.","jevDecisionPointJa":"条件を yes/no 確率質問へ変え、ローカル閾値の判定と確率を返す。","jevDecisionPointKo":"조건을 yes/no 확률 질문으로 바꾸고 로컬 임곗값 결과와 확률을 반환한다.","highlightBenefitJa":"タスク結果の確認用で、メッセージシステム全体を Jev が制御するわけではない。","highlightBenefitKo":"작업 결과를 구조적으로 확인하며 전체 메시지 시스템을 Jev가 제어하는 것은 아니다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/ldbumble/taskuary/blob/4ad29d7b292a7899767338cfcc83b2dde8f43330/README.md","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/ldbumble/taskuary/blob/4ad29d7b292a7899767338cfcc83b2dde8f43330/README.zh-CN.md","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/ldbumble/taskuary/blob/4ad29d7b292a7899767338cfcc83b2dde8f43330/taskuary/jev.py","note":"固定版本说明与源码；本站未执行该项目。"}],"pinned":false,"topics":["agentic-workflow","ai","ai-agents","automation","claude","claude-code","codex","email-triage","fastapi","kanban","llm","local-first","mcp","microsoft-teams","outlook","productivity","python","react","slack","task-management"]},{"id":"yonatangross:orchestkit","name":"orchestkit","author":"yonatangross","url":"https://github.com/yonatangross/orchestkit","category":"CLI & Pipelines","plainSummary":"OrchestKit 可选用 Jev 给编程会话分类，符合阈值时用结果决定会话颜色。","plainSummaryEn":"OrchestKit can optionally use Jev to classify coding sessions and set their colors when confidence meets a threshold.","jevDecisionPoint":"将首条任务提示与分支状态分类为开发、排错等工作类型；本地规则选择采用或回退。","jevDecisionPointEn":"Classifies the first task prompt and branch state by work type; local policy accepts the result or falls back.","highlightBenefit":"用工作类型区分会话，同时保留 shadow 对照模式。","highlightBenefitEn":"Distinguishes sessions by work type and supports a shadow comparison mode.","tags":["coding-agents","classification-ranking"],"stars":280,"forks":34,"openIssues":75,"license":"MIT","language":"TypeScript","lastCommitAt":"2026-09-21T21:29:12Z","createdAt":"2025-12-31T05:58:52Z","metadataFetchedAt":"2026-09-21T21:59:26.939Z","avatarUrl":"https://avatars.githubusercontent.com/u/19869074?v=4","summarySource":"human-reviewed","claimStatus":"Jev 和 shadow 模式均需主动配置；shadow 只记录不应用，不能说整个 OrchestKit 由 Jev 决策。","claimStatusEn":"Jev and shadow modes require explicit configuration. Shadow logs decisions without applying them; this does not mean Jev drives all of OrchestKit.","plainSummaryJa":"OrchestKit は任意で Jev にコーディングセッションを分類させ、閾値を満たすと表示色に使う。","plainSummaryKo":"OrchestKit은 선택적으로 Jev로 코딩 세션을 분류하고 임계값을 충족하면 표시 색상에 반영한다.","jevDecisionPointJa":"最初のタスクとブランチ状態から作業種別を選び、ローカルルールが採用か代替処理を決める。","jevDecisionPointKo":"첫 작업 프롬프트와 브랜치 상태로 작업 유형을 선택하고 로컬 규칙이 채택이나 대체 처리를 결정한다.","highlightBenefitJa":"作業種別でセッションを区別し、shadow 比較モードも使える。","highlightBenefitKo":"작업 유형으로 세션을 구별하고 shadow 비교 모드도 사용할 수 있다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"confirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/yonatangross/orchestkit/blob/569095f35ce61a4881197dfb8117f4c3312e2bee/src/hooks/src/lib/session-category-provider.ts","note":"Fixed-commit source reviewed / 已核对固定版本源码"},{"url":"https://github.com/yonatangross/orchestkit/blob/569095f35ce61a4881197dfb8117f4c3312e2bee/README.md","note":"Fixed-commit source reviewed / 已核对固定版本源码"}],"pinned":false,"topics":["agent-orchestration","agents","ai-agents","ai-development","anthropic","claude-code","claude-code-plugin","claude-plugin","developer-tools","fastapi","langgraph","llm","mcp","rag","react","security","testing","typescript"]},{"id":"kitze:skillbox","name":"skillbox","author":"kitze","url":"https://github.com/kitze/skillbox","category":"Routing & Cost Optimization","plainSummary":"自建一个有版本管理的 Agent 技能库，还能选配 Jev 推荐：告诉它当前任务，从你有权限使用的技能里挑更相关的。","plainSummaryEn":"A self-hosted, versioned Agent skills library with optional Jev recommendations.","jevDecisionPoint":"对候选技能逐项评估任务相关度，再由应用整理推荐结果。","jevDecisionPointEn":"Jev scores task relevance among skills the client is allowed to access.","highlightBenefit":"技能查找不只依赖关键词；未配置或调用失败时仍可使用确定性搜索。","highlightBenefitEn":"Suggests relevant skills while retaining the library’s client access scope.","tags":["llm-routing-cost","mcp-integrations"],"stars":223,"forks":19,"openIssues":4,"license":"MIT","language":"TypeScript","lastCommitAt":"2026-09-19T21:46:18Z","createdAt":"2026-09-17T17:23:37Z","metadataFetchedAt":"2026-09-21T21:59:29.924Z","avatarUrl":"https://avatars.githubusercontent.com/u/1160594?v=4","summarySource":"human-reviewed","claimStatus":"Jev 用于可选技能推荐，不是日志裁剪器。","claimStatusEn":"Jev provides optional skill recommendations, not log pruning; recommendation accuracy was not measured here.","plainSummaryJa":"任意の Jev 推薦機能を持つ、自前ホスト型の版管理付き Agent skill ライブラリー。","plainSummaryKo":"선택형 Jev 추천을 제공하는 자체 호스팅·버전 관리 Agent skill 라이브러리.","jevDecisionPointJa":"クライアントが利用できる skill の中で、タスクとの関連性を評価する。","jevDecisionPointKo":"클라이언트가 접근 가능한 skill 중 작업과의 관련성을 평가한다.","highlightBenefitJa":"アクセス範囲を保ったまま、関連 skill を探せる。","highlightBenefitKo":"접근 범위를 유지하면서 관련 skill을 찾는다.","claimStatusJa":"固定バージョンの文書とソースを確認。実行・性能・安全性は当サイトで独立検証していません。","claimStatusKo":"고정 버전 문서와 소스를 확인했습니다. 실행·성능·안전성은 이 사이트에서 독립적으로 검증하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/kitze/skillbox/blob/d83ba4ecd254c8dfa6a759d1feb5141384e26a9e/README.md","note":"README.md"},{"url":"https://github.com/kitze/skillbox/blob/d83ba4ecd254c8dfa6a759d1feb5141384e26a9e/src/server/recommendations.ts","note":"src/server/recommendations.ts"}],"pinned":false},{"id":"quentincody:interlinked-cli","name":"interlinked-cli","author":"QuentinCody","url":"https://github.com/QuentinCody/interlinked-cli","category":"Security & Guardrails","plainSummary":"Interlinked 在编程 Agent 的本地检查之外，提供可选 Jev 判断与证据检查。","plainSummaryEn":"Interlinked adds optional Jev judgments and evidence checks to local coding-agent checks.","jevDecisionPoint":"评估测试名称、文档声明和对应证据等输入，给规则流程补充语义判断。","jevDecisionPointEn":"Assesses inputs such as test names, documentation claims and supporting evidence.","highlightBenefit":"把模型建议与确定性规则分开，保留检查记录。","highlightBenefitEn":"Keeps model advice distinct from deterministic rules and records checks.","tags":["coding-agents","security-guardrails","cli-git-gates"],"stars":178,"forks":7,"openIssues":4,"license":"MIT","lastCommitAt":"2026-09-17T13:50:34Z","createdAt":"2026-04-23T20:13:07Z","metadataFetchedAt":"2026-09-21T21:59:32.949Z","avatarUrl":"https://avatars.githubusercontent.com/u/33259999?v=4","summarySource":"human-reviewed","claimStatus":"已核对固定版本 README 与集成源码。","claimStatusEn":"README and integration source reviewed at a fixed commit; not independently run or benchmarked by this site.","plainSummaryJa":"Interlinked はローカルの Agent 検査に、任意の Jev 判断と根拠確認を加える。","plainSummaryKo":"Interlinked는 로컬 코딩 Agent 검사에 선택형 Jev 판단과 근거 확인을 더한다.","jevDecisionPointJa":"テスト名、文書の主張、対応する根拠などを評価する。","jevDecisionPointKo":"테스트 이름, 문서 주장과 대응 근거 등을 평가한다.","highlightBenefitJa":"モデルの助言を決定的な規則と分けて記録する。","highlightBenefitKo":"모델 조언을 결정적 규칙과 구분하고 검사 기록을 남긴다.","claimStatusJa":"固定版の README と実装を確認済み。このサイトでは独立した実行・性能検証をしていません。","claimStatusKo":"고정 버전 README와 구현을 확인했습니다. 이 사이트에서 독립적인 실행·성능 검증은 하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/QuentinCody/interlinked-cli/blob/207330d8131c5203ecc74e9fd4c24ba416463718/src/harness/jev/client.ts#L14","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/QuentinCody/interlinked-cli/blob/207330d8131c5203ecc74e9fd4c24ba416463718/README.md","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/QuentinCody/interlinked-cli/blob/207330d8131c5203ecc74e9fd4c24ba416463718/docs/README.md","note":"固定版本的实现与说明 / Pinned source and documentation"}],"pinned":false},{"id":"danieljvdm:effect-agent","name":"effect-agent","author":"danieljvdm","url":"https://github.com/danieljvdm/effect-agent","category":"SDK & Decision Frameworks","plainSummary":"Effect Agent 的 TypeSafe 决策 provider，支持类型化问题集与可选模型选择。","plainSummaryEn":"An Effect Agent TypeSafe decision provider for typed question sets and optional model selection.","jevDecisionPoint":"Jev 回答概率、选择和评分问题，应用按结果执行状态转换或选择模型。","jevDecisionPointEn":"Jev answers probability, choice and score questions for application transitions or model selection.","highlightBenefit":"将判断纳入 Effect 的依赖与错误处理流程，重试和超时由应用配置。","highlightBenefitEn":"Integrates decisions into Effect dependencies and errors; applications configure retries and deadlines.","tags":["multilanguage-sdk","typed-decisions"],"stars":121,"forks":6,"openIssues":10,"license":"MIT","language":"TypeScript","lastCommitAt":"2026-09-21T21:01:38Z","createdAt":"2026-07-29T22:08:15Z","metadataFetchedAt":"2026-09-21T21:59:35.945Z","avatarUrl":"https://avatars.githubusercontent.com/u/3450486?v=4","summarySource":"human-reviewed","claimStatus":"已核对固定版本 README 与调用源码。","claimStatusEn":"Reviewed the fixed-version README and call path.","plainSummaryJa":"型付き質問集合と任意のモデル選択を扱う Effect Agent の TypeSafe provider。","plainSummaryKo":"구조화된 질문 집합과 선택적 모델 선택을 지원하는 Effect Agent TypeSafe provider.","jevDecisionPointJa":"Jev の確率・選択・採点を、状態遷移やモデル選択に利用する。","jevDecisionPointKo":"Jev의 확률·선택·채점을 상태 전환이나 모델 선택에 사용한다.","highlightBenefitJa":"Effect の依存・エラー処理に統合し、再試行と期限はアプリが設定する。","highlightBenefitKo":"Effect 의존성·오류 처리에 통합하며 재시도와 기한은 앱에서 설정한다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/danieljvdm/effect-agent/blob/88005e497e9b627eeb16d670f278903c57601da9/README.md","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/danieljvdm/effect-agent/blob/88005e497e9b627eeb16d670f278903c57601da9/packages/ai-typesafe/README.md","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/danieljvdm/effect-agent/blob/88005e497e9b627eeb16d670f278903c57601da9/packages/ai-typesafe/src/TypeSafeClient.ts","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/danieljvdm/effect-agent/blob/88005e497e9b627eeb16d670f278903c57601da9/packages/ai-decision/src/AutoModel.ts","note":"固定版本说明与源码；本站未执行该项目。"}],"pinned":false},{"id":"devmortimer:pi-warden","name":"pi-warden","author":"DevMortimer","url":"https://github.com/DevMortimer/pi-warden","category":"Security & Guardrails","plainSummary":"给 Pi Agent 加项目规则、越界操作、重复失败和完成声明的检查。","plainSummaryEn":"Adds checks for project rules, out-of-scope actions, repeated failures, and completion claims to Pi Agents.","jevDecisionPoint":"用 Jev 判断写入是否违反规则、操作是否不可逆或偏离任务，并把结果反馈到 Agent。","jevDecisionPointEn":"Uses Jev to judge rule violations, irreversible actions, and task drift, feeding results back to the Agent.","highlightBenefit":"把规则问题和需要确认的操作带回当前工作流。","highlightBenefitEn":"Brings rule findings and actions needing confirmation into the active workflow.","tags":["security-guardrails","coding-agents"],"stars":127,"forks":12,"openIssues":3,"license":"MIT","language":"TypeScript","lastCommitAt":"2026-09-21T17:39:34Z","createdAt":"2026-09-16T03:58:46Z","metadataFetchedAt":"2026-09-21T21:59:38.942Z","avatarUrl":"https://avatars.githubusercontent.com/u/161127243?v=4","summarySource":"human-reviewed","claimStatus":"源码包含接入及操作检查；作者的有限样本对照不等于安全保证，本站未独立验证防护或降本效果。","claimStatusEn":"Source contains integration and action checks. The author’s limited comparisons are not a security guarantee; this site has not independently verified protection or savings.","plainSummaryJa":"Pi Agent にプロジェクトルール、範囲外の操作、反復失敗、完了宣言のチェックを追加する。","plainSummaryKo":"Pi Agent에 프로젝트 규칙, 범위 밖 작업, 반복 실패, 완료 선언 검사를 추가한다.","jevDecisionPointJa":"ルール違反、不可逆操作、タスクの逸脱を Jev が判断し Agent に返す。","jevDecisionPointKo":"규칙 위반, 되돌릴 수 없는 작업, 작업 이탈을 Jev가 판단해 Agent에 돌려준다.","highlightBenefitJa":"ルール上の問題と確認が必要な操作を現在の作業へ戻す。","highlightBenefitKo":"규칙 문제와 확인이 필요한 작업을 현재 워크플로에 전달한다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"confirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/DevMortimer/pi-warden/blob/e6c801679464b1a9624225114eb5fb443c27d823/src/guard.ts","note":"Fixed-commit source reviewed / 已核对固定版本源码"},{"url":"https://github.com/DevMortimer/pi-warden/blob/e6c801679464b1a9624225114eb5fb443c27d823/README.md","note":"Fixed-commit source reviewed / 已核对固定版本源码"}],"pinned":false,"topics":["guardrails","pi-extension","pi-package","typesafe"]},{"id":"pithings:advocaat","name":"advocaat","author":"pithings","url":"https://github.com/pithings/advocaat","category":"SDK & Decision Frameworks","plainSummary":"用简短的 TypeScript 调用向 Jev 提问。把同一份数据里的多个判断一次写好，直接拿到概率、选项和分数。","plainSummaryEn":"A small TypeScript client for asking Jev multiple typed questions about the same data.","jevDecisionPoint":"将带类型的问题转成 System One 请求，并把命名答案映射回调用方。","jevDecisionPointEn":"Combines yes/no, Choice and Score questions and maps responses to typed values.","highlightBenefit":"减少手写请求和解析代码，也支持通过 Vercel 网关调用。","highlightBenefitEn":"Reuses one interface for probabilities, choices and scores.","tags":["multilanguage-sdk","typed-decisions"],"stars":89,"forks":1,"openIssues":0,"license":"MIT","language":"TypeScript","lastCommitAt":"2026-09-18T17:55:47Z","createdAt":"2026-09-16T21:23:24Z","metadataFetchedAt":"2026-09-21T21:59:41.957Z","avatarUrl":"https://avatars.githubusercontent.com/u/263128954?v=4","summarySource":"human-reviewed","claimStatus":"已核对客户端接口与请求实现；类型封装不保证模型判断正确。","claimStatusEn":"Client interfaces and request code were inspected; type wrappers do not guarantee correct judgments. Not run here.","plainSummaryJa":"同じデータに複数の型付き質問を送る、小さな TypeScript 製 Jev クライアント。","plainSummaryKo":"같은 데이터에 여러 타입 질문을 보내는 작은 TypeScript Jev 클라이언트.","jevDecisionPointJa":"Yes/No・Choice・Score をまとめ、回答を型付き値へ変換する。","jevDecisionPointKo":"예/아니요·Choice·Score 질문을 묶고 답을 타입 값에 매핑한다.","highlightBenefitJa":"確率・選択・評価に共通インターフェースを使える。","highlightBenefitKo":"확률·선택·평가에 공통 인터페이스를 사용한다.","claimStatusJa":"固定バージョンの文書とソースを確認。実行・性能・安全性は当サイトで独立検証していません。","claimStatusKo":"고정 버전 문서와 소스를 확인했습니다. 실행·성능·안전성은 이 사이트에서 독립적으로 검증하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/pithings/advocaat/blob/bc46287fc1102b95852a81d679c6e34a2c44f4a2/README.md","note":"README.md"},{"url":"https://github.com/pithings/advocaat/blob/bc46287fc1102b95852a81d679c6e34a2c44f4a2/src/api.ts","note":"src/api.ts"}],"pinned":false},{"id":"getsynkora:synkora-ai","name":"synkora-ai","author":"getsynkora","url":"https://github.com/getsynkora/synkora-ai","category":"MCP & Integrations","plainSummary":"Synkora Agent 平台内置可选 TypeSafe 客户端与工具，用于分类、评分和是非判断。","plainSummaryEn":"Synkora includes optional TypeSafe client tools for classification, scoring and yes/no judgments.","jevDecisionPoint":"把工作流状态和命名问题发送给 Jev，并将结构化答案交还 Agent。","jevDecisionPointEn":"Sends workflow state and named questions to Jev and returns structured answers to agents.","highlightBenefit":"让现有平台工作流调用同一组判断能力。","highlightBenefitEn":"Makes these judgments available within existing platform workflows.","tags":["mcp-integrations","classification-ranking","typed-decisions"],"stars":35,"forks":6,"openIssues":4,"license":"MIT","lastCommitAt":"2026-09-20T07:12:37Z","createdAt":"2026-03-22T01:24:44Z","metadataFetchedAt":"2026-09-21T21:59:44.980Z","avatarUrl":"https://avatars.githubusercontent.com/u/271684909?v=4","summarySource":"human-reviewed","claimStatus":"已核对固定版本 README 与集成源码。","claimStatusEn":"README and integration source reviewed at a fixed commit; not independently run or benchmarked by this site.","plainSummaryJa":"Synkora は分類、採点、真偽判断向けの任意の TypeSafe ツールを備える。","plainSummaryKo":"Synkora는 분류, 점수와 참·거짓 판단을 위한 선택형 TypeSafe 도구를 제공한다.","jevDecisionPointJa":"処理の状態と名前付き質問を Jev に送り、回答を Agent に返す。","jevDecisionPointKo":"작업 상태와 이름 붙인 질문을 Jev에 보내 Agent에 답을 돌려준다.","highlightBenefitJa":"既存のプラットフォーム処理から判断機能を使える。","highlightBenefitKo":"기존 플랫폼 흐름에서 판단 기능을 사용할 수 있다.","claimStatusJa":"固定版の README と実装を確認済み。このサイトでは独立した実行・性能検証をしていません。","claimStatusKo":"고정 버전 README와 구현을 확인했습니다. 이 사이트에서 독립적인 실행·성능 검증은 하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/getsynkora/synkora-ai/blob/2ef7600c9772a025110e49ee06e66e18bb300843/api/src/core/typesafe_client.py#L2","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/getsynkora/synkora-ai/blob/2ef7600c9772a025110e49ee06e66e18bb300843/README.md","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/getsynkora/synkora-ai/blob/2ef7600c9772a025110e49ee06e66e18bb300843/api/README.md","note":"固定版本的实现与说明 / Pinned source and documentation"}],"pinned":false},{"id":"hackclub:ai","name":"ai","author":"hackclub","url":"https://github.com/hackclub/ai","category":"SDK & Decision Frameworks","plainSummary":"Hack Club AI 代理中的 Jev 转发接口，复用已有鉴权、限额和用量记录。","plainSummaryEn":"A Jev forwarding endpoint in the Hack Club AI proxy, using its authentication, limits and usage logging.","jevDecisionPoint":"把获准用户的结构化请求转发到 TypeSafe，返回模型结果并记录用量。","jevDecisionPointEn":"Forwards authorized structured requests to TypeSafe and returns answers with usage accounting.","highlightBenefit":"为现有代理增加决策 API 入口，不是 MCP 工具服务器。","highlightBenefitEn":"Adds a decision API to the existing proxy rather than an MCP tool server.","tags":["multilanguage-sdk","typed-decisions"],"stars":133,"forks":54,"openIssues":27,"license":null,"language":"TypeScript","lastCommitAt":"2026-09-17T21:48:35Z","createdAt":"2025-01-25T21:57:58Z","metadataFetchedAt":"2026-09-21T21:59:47.958Z","avatarUrl":"https://avatars.githubusercontent.com/u/5633654?v=4","summarySource":"human-reviewed","claimStatus":"已核对固定版本 README 与调用源码。","claimStatusEn":"Reviewed the fixed-version README and call path.","plainSummaryJa":"Hack Club AI プロキシの認証・制限・使用量記録を使う Jev 転送エンドポイント。","plainSummaryKo":"Hack Club AI 프록시의 인증·한도·사용량 기록을 활용하는 Jev 전달 엔드포인트.","jevDecisionPointJa":"許可された構造化要求を TypeSafe へ転送し、回答と使用量を扱う。","jevDecisionPointKo":"허가된 구조화 요청을 TypeSafe에 전달하고 응답과 사용량을 처리한다.","highlightBenefitJa":"既存プロキシに判断 API を加えるもので、MCP サーバーではない。","highlightBenefitKo":"기존 프록시에 판단 API를 추가하며 MCP 서버는 아니다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"unconfirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/hackclub/ai/blob/a76ea2cb159f707a60107935a5b2e0dbdc7455f5/README.md","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/hackclub/ai/blob/a76ea2cb159f707a60107935a5b2e0dbdc7455f5/src/routes/proxy/v1/jev.ts","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/hackclub/ai/blob/a76ea2cb159f707a60107935a5b2e0dbdc7455f5/src/routes/proxy/v1/index.ts","note":"固定版本说明与源码；本站未执行该项目。"}],"pinned":false},{"id":"kitze:unclutter","name":"unclutter","author":"kitze","url":"https://github.com/kitze/unclutter","category":"Security & Guardrails","plainSummary":"用 Jev 帮浏览器扩展识别网页中的广告、促销与订阅弹窗，并保存可复用的隐藏规则。","plainSummaryEn":"A browser extension using Jev to identify ads, promotions, and subscription popups and save reusable hiding rules.","jevDecisionPoint":"对候选页面元素做结构化判断，再由扩展应用本地隐藏规则。","jevDecisionPointEn":"Evaluates candidate page elements before the extension applies local hiding rules.","highlightBenefit":"把一次页面判断转成可重复使用的规则。","highlightBenefitEn":"Turns page judgments into reusable rules.","tags":["browser-automation","classification-ranking"],"stars":168,"forks":13,"openIssues":1,"license":"MIT","language":"TypeScript","lastCommitAt":"2026-09-18T08:12:00Z","createdAt":"2026-09-17T11:58:56Z","metadataFetchedAt":"2026-09-21T21:59:50.919Z","avatarUrl":"https://avatars.githubusercontent.com/u/1160594?v=4","summarySource":"human-reviewed","claimStatus":"已核对固定版本的 README 与接入代码。","claimStatusEn":"Reviewed the README and integration code at a fixed commit. This site has not run the project or independently verified performance, security, or business outcomes.","plainSummaryJa":"広告、販促、購読ポップアップを Jev で判定し、再利用可能な非表示ルールを保存する拡張。","plainSummaryKo":"광고, 홍보, 구독 팝업을 Jev로 판단해 재사용 가능한 숨김 규칙을 저장하는 확장.","jevDecisionPointJa":"候補のページ要素を評価し、拡張がローカルの非表示ルールを適用する。","jevDecisionPointKo":"후보 페이지 요소를 평가하면 확장이 로컬 숨김 규칙을 적용한다.","highlightBenefitJa":"ページ上の判断を再利用できるルールにする。","highlightBenefitKo":"페이지 판단을 재사용 가능한 규칙으로 만든다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"confirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/kitze/unclutter/blob/9ef9beccc1e57b4e3115ae68644b8fc9c19c29f6/lib/jev.ts","note":"Fixed-commit source reviewed / 已核对固定版本源码"},{"url":"https://github.com/kitze/unclutter/blob/9ef9beccc1e57b4e3115ae68644b8fc9c19c29f6/README.md","note":"Fixed-commit source reviewed / 已核对固定版本源码"}],"pinned":false},{"id":"cequence-io:openai-scala-client","name":"openai-scala-client","author":"cequence-io","url":"https://github.com/cequence-io/openai-scala-client","category":"SDK & Decision Frameworks","plainSummary":"让 Scala 应用也能接入 Jev。这个多模型客户端新增了独立 TypeSafe 模块，用状态和判断题获取结构化答案。","plainSummaryEn":"A dedicated TypeSafe module in a Scala client that supports multiple AI providers.","jevDecisionPoint":"调用 System One 返回单选、评分和是非概率；可把受支持的封闭 JSON Schema 转成判断题。","jevDecisionPointEn":"Sends shared state and typed questions to Jev and parses structured answers.","highlightBenefit":"能沿用 Scala 的异步接口和错误处理；Jev 是额外支持的 provider，不承担普通聊天生成。","highlightBenefitEn":"Lets Scala applications reuse provider transport and error-handling conventions.","tags":["multilanguage-sdk","typed-decisions"],"stars":248,"forks":39,"openIssues":9,"license":"MIT","language":"Scala","lastCommitAt":"2026-09-18T18:14:56Z","createdAt":"2023-01-25T10:09:04Z","metadataFetchedAt":"2026-09-21T21:59:54.001Z","avatarUrl":"https://avatars.githubusercontent.com/u/122438709?v=4","summarySource":"human-reviewed","claimStatus":"核对的是独立 TypeSafe 模块；不能把整个多模型仓库的采用度等同于 Jev 采用度。","claimStatusEn":"This covers the dedicated TypeSafe module; adoption of the overall multi-provider repository is not Jev adoption. Not run here.","plainSummaryJa":"複数の AI provider に対応する Scala クライアント内の TypeSafe 専用モジュール。","plainSummaryKo":"여러 AI provider를 지원하는 Scala 클라이언트의 TypeSafe 전용 모듈.","jevDecisionPointJa":"共有状態と型付き質問を Jev に送り、構造化回答を解析する。","jevDecisionPointKo":"공유 상태와 타입 질문을 Jev에 보내 구조화된 답을 해석한다.","highlightBenefitJa":"Scala アプリで通信とエラー処理の仕組みを再利用できる。","highlightBenefitKo":"Scala 앱에서 통신과 오류 처리 방식을 재사용한다.","claimStatusJa":"固定バージョンの文書とソースを確認。実行・性能・安全性は当サイトで独立検証していません。","claimStatusKo":"고정 버전 문서와 소스를 확인했습니다. 실행·성능·안전성은 이 사이트에서 독립적으로 검증하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/cequence-io/openai-scala-client/blob/cfabe8842344da713615d0121a7804840649eb46/README.md","note":"README.md"},{"url":"https://github.com/cequence-io/openai-scala-client/blob/cfabe8842344da713615d0121a7804840649eb46/typesafe-client/src/main/scala/io/cequence/openaiscala/typesafe/service/impl/TypeSafeServiceImpl.scala","note":"typesafe-client/src/main/scala/io/cequence/openaiscala/typesafe/service/impl/TypeSafeServiceImpl.scala"}],"pinned":false,"topics":["anthropic","anthropic-api","aws-bedrock","chatgpt","gemini","gemini-ai","groq-api","llms","nlp-library","openai","openai-api-client","perplexity-api","scala","vertex-ai-gemini-api","vlms"]},{"id":"phd-peter:ego-jev","name":"ego-jev","author":"phd-peter","url":"https://github.com/phd-peter/ego-jev","category":"Browser & OS Action","plainSummary":"把 Ego Lite 的浏览器快照与操作接到一个有步骤上限的 Jev 决策循环。","plainSummaryEn":"Connects Ego Lite snapshots and browser actions to a bounded Jev decision loop.","jevDecisionPoint":"只从当前快照的候选元素和支持动作中选择；输入文字时可调用另一个模型。","jevDecisionPointEn":"Selects only current snapshot candidates and supported actions; another model can supply typed text.","highlightBenefit":"用当前快照的引用执行动作，并记录每步状态。","highlightBenefitEn":"Executes against current snapshot references and records step state.","tags":["browser-automation","typed-decisions"],"stars":0,"forks":0,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-18T07:04:16Z","createdAt":"2026-09-18T06:51:40Z","metadataFetchedAt":"2026-09-21T21:59:56.980Z","avatarUrl":"https://avatars.githubusercontent.com/u/175372403?v=4","summarySource":"human-reviewed","claimStatus":"已核对固定版本 README 与集成源码。","claimStatusEn":"README and integration source reviewed at a fixed commit; not independently run or benchmarked by this site.","plainSummaryJa":"Ego Lite のスナップショットと操作を、回数制限のある Jev 判断ループにつなぐ。","plainSummaryKo":"Ego Lite 스냅샷과 브라우저 동작을 횟수가 제한된 Jev 판단 루프에 연결한다.","jevDecisionPointJa":"現在のスナップショットの要素と対応操作から選び、必要な入力文は別モデルで補う。","jevDecisionPointKo":"현재 스냅샷 후보와 지원 동작에서만 선택하며 입력 문장은 별도 모델이 제공할 수 있다.","highlightBenefitJa":"現在の参照で操作し、各段階の状態を記録する。","highlightBenefitKo":"현재 스냅샷 참조로 실행하고 단계별 상태를 기록한다.","claimStatusJa":"固定版の README と実装を確認済み。このサイトでは独立した実行・性能検証をしていません。","claimStatusKo":"고정 버전 README와 구현을 확인했습니다. 이 사이트에서 독립적인 실행·성능 검증은 하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/phd-peter/ego-jev/blob/0da45e4f3a3bdad56402b5c38226ad3a2e10409c/src/jev-client.ts#L1","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/phd-peter/ego-jev/blob/0da45e4f3a3bdad56402b5c38226ad3a2e10409c/README.md","note":"固定版本的实现与说明 / Pinned source and documentation"}],"pinned":false},{"id":"rokcso:bluenoise","name":"bluenoise","author":"rokcso","url":"https://github.com/rokcso/bluenoise","category":"Context GC & Filter","plainSummary":"为 X/Twitter 过滤帖子与回复的浏览器扩展；默认用本地规则，可选择用 Jev 检查未匹配的回复。","plainSummaryEn":"An X/Twitter filtering extension using local rules by default, with optional Jev checks for unmatched replies.","jevDecisionPoint":"启用实验性 AI 后，将未命中规则的回复交给 Jev 评估，再由本地阈值决定是否隐藏。","jevDecisionPointEn":"When experimental AI is enabled, Jev evaluates replies missed by rules and local thresholds decide whether to hide them.","highlightBenefit":"先应用可逆的本地规则，只对需要的内容追加模型判断。","highlightBenefitEn":"Applies reversible local rules first and adds model judgment only when requested.","tags":["browser-automation","classification-ranking"],"stars":90,"forks":5,"openIssues":3,"license":"MIT","language":"TypeScript","lastCommitAt":"2026-09-18T08:57:06Z","createdAt":"2026-08-20T10:08:40Z","metadataFetchedAt":"2026-09-21T21:59:59.960Z","avatarUrl":"https://avatars.githubusercontent.com/u/47956905?v=4","summarySource":"human-reviewed","claimStatus":"Jev 功能默认关闭；开启后相关回复会发送到 TypeSafe，不能把 AI 模式描述成完全本地处理。","claimStatusEn":"Jev is disabled by default. Enabling it sends relevant replies to TypeSafe, so AI mode is not fully local processing. This site has not run it.","plainSummaryJa":"X/Twitter の投稿や返信をローカルルールで整理し、未一致の返信だけ任意で Jev に確認させる拡張。","plainSummaryKo":"X/Twitter 게시물과 답글을 로컬 규칙으로 정리하고 미일치 답글만 선택적으로 Jev에 확인하는 확장.","jevDecisionPointJa":"実験的 AI を有効にすると、ルールに該当しない返信を評価し、閾値で非表示を決める。","jevDecisionPointKo":"실험적 AI를 켜면 규칙에 해당하지 않는 답글을 평가하고 임계값으로 숨김을 결정한다.","highlightBenefitJa":"可逆なローカルルールを先に適用し、必要に応じてモデル判断を追加する。","highlightBenefitKo":"되돌릴 수 있는 로컬 규칙을 먼저 적용하고 필요할 때 모델 판단을 추가한다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"confirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/rokcso/bluenoise/blob/ef81ea7a7c3677501d6de8f9235a4d6a866b573a/entrypoints/background.ts","note":"Fixed-commit source reviewed / 已核对固定版本源码"},{"url":"https://github.com/rokcso/bluenoise/blob/ef81ea7a7c3677501d6de8f9235a4d6a866b573a/README.md","note":"Fixed-commit source reviewed / 已核对固定版本源码"}],"pinned":false,"topics":["chrome-extension","extension","twitter"]},{"id":"xiechengyuan:jev-gomoku","name":"jev-gomoku","author":"XieChengYuan","url":"https://github.com/XieChengYuan/jev-gomoku","category":"Evaluation & Observability","plainSummary":"弈瞬：同时运行九盘 15×15 五子棋，让两位 Jev 玩家比较不同输入信息，并逐手检查请求与返回。","plainSummaryEn":"A nine-board, 15×15 Gomoku workbench comparing how two Jev players respond to different input representations.","jevDecisionPoint":"从本地规则生成的候选落点中做 Choice 判断，对比棋盘、战术事实等不同输入。","jevDecisionPointEn":"Jev chooses from locally generated move candidates while the experiment varies board and tactical information.","highlightBenefit":"支持带明确标记的历史回放和自带密钥的实时对战，保留逐手实验记录。","highlightBenefitEn":"Supports labeled replays and bring-your-own-key live games with per-move records.","tags":["games-simulation","evaluation-benchmarks"],"stars":1,"forks":0,"openIssues":0,"license":null,"language":"JavaScript","lastCommitAt":"2026-09-18T08:38:09Z","createdAt":"2026-09-18T08:38:10Z","metadataFetchedAt":"2026-09-21T22:00:02.987Z","avatarUrl":"https://avatars.githubusercontent.com/u/33199973?v=4","summarySource":"human-reviewed","claimStatus":"免费模式播放保存的对局，不产生新推理；对局胜负不能证明通用棋力。本站未重跑。","claimStatusEn":"Free mode replays saved games without new inference. Match outcomes do not establish general playing strength; not rerun here.","plainSummaryJa":"九つの 15×15 五目並べ盤で、二人の Jev に渡す情報の違いを比較する実験台。","plainSummaryKo":"9개의 15×15 오목판에서 두 Jev 플레이어에게 제공하는 정보 차이를 비교하는 실험 도구.","jevDecisionPointJa":"ローカル生成の着手候補から選び、盤面情報や戦術情報の条件を変える。","jevDecisionPointKo":"로컬에서 생성한 착수 후보 중 선택하며 보드·전술 정보 조건을 바꾼다.","highlightBenefitJa":"明示されたリプレイと自分のキーによる対局で、一手ごとの記録を確認できる。","highlightBenefitKo":"표시된 리플레이와 개인 키를 쓰는 실시간 대국에서 수별 기록을 확인한다.","claimStatusJa":"固定バージョンの文書とソースを確認。実行・性能・安全性は当サイトで独立検証していません。","claimStatusKo":"고정 버전 문서와 소스를 확인했습니다. 실행·성능·안전성은 이 사이트에서 독립적으로 검증하지 않았습니다.","licenseStatus":"unconfirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/XieChengYuan/jev-gomoku/blob/2aac72cd54ab63fd35e7d29e65def787dda2b333/README.md","note":"README.md"},{"url":"https://github.com/XieChengYuan/jev-gomoku/blob/2aac72cd54ab63fd35e7d29e65def787dda2b333/src/decision.js","note":"src/decision.js"},{"url":"https://github.com/XieChengYuan/jev-gomoku/blob/2aac72cd54ab63fd35e7d29e65def787dda2b333/src/online.js","note":"src/online.js"}],"pinned":false},{"id":"moritzkremb:jev-voice-browser","name":"jev-voice-browser","author":"moritzkremb","url":"https://github.com/moritzkremb/jev-voice-browser","category":"Browser & OS Action","plainSummary":"用语音控制 Playwright 浏览器，将逐步转写的口令交给 Jev 判断。","plainSummaryEn":"Controls a Playwright browser by sending incremental speech transcripts to Jev.","jevDecisionPoint":"选择意图、元素、网址或原文片段，并判断口令是否完整、是否涉及敏感动作。","jevDecisionPointEn":"Chooses intent, elements, URLs or verbatim spans and judges command completeness and sensitive actions.","highlightBenefit":"界面显示概率、动作与请求时间，便于观察语音交互。","highlightBenefitEn":"Shows probabilities, actions and request timing during voice interaction.","tags":["browser-automation","voice-conversation"],"stars":203,"forks":27,"openIssues":2,"license":"MIT","lastCommitAt":"2026-09-21T08:57:54Z","createdAt":"2026-09-17T22:40:19Z","metadataFetchedAt":"2026-09-21T22:00:05.954Z","avatarUrl":"https://avatars.githubusercontent.com/u/33087790?v=4","summarySource":"human-reviewed","claimStatus":"约 250–350 ms 是作者报告的单次请求时延，不是每条语音指令的端到端保证。 已核对固定版本 README 与集成源码。","claimStatusEn":"The reported 250–350 ms measures an individual request, not guaranteed end-to-end response time for every voice command. README and integration source reviewed at a fixed commit; not independently run or benchmarked by this site.","plainSummaryJa":"音声の逐次書き起こしを Jev に送り、Playwright ブラウザを操作する。","plainSummaryKo":"점진적인 음성 전사를 Jev에 보내 Playwright 브라우저를 조작한다.","jevDecisionPointJa":"意図、要素、URL、原文範囲を選び、命令の完結性や注意を要する操作を判断する。","jevDecisionPointKo":"의도, 요소, URL, 원문 구간을 고르고 명령 완결성과 민감한 동작을 판단한다.","highlightBenefitJa":"音声操作中の確率、動作、リクエスト時間を表示する。","highlightBenefitKo":"음성 조작 중 확률, 동작과 요청 시간을 표시한다.","claimStatusJa":"固定版の README と実装を確認済み。このサイトでは独立した実行・性能検証をしていません。","claimStatusKo":"고정 버전 README와 구현을 확인했습니다. 이 사이트에서 독립적인 실행·성능 검증은 하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/moritzkremb/jev-voice-browser/blob/054db0f3dbf537af63a8117632d3f941ccd520e1/src/jev.js#L123","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/moritzkremb/jev-voice-browser/blob/054db0f3dbf537af63a8117632d3f941ccd520e1/README.md","note":"固定版本的实现与说明 / Pinned source and documentation"}],"pinned":false},{"id":"yusukebe:hono-jev-router","name":"hono-jev-router","author":"yusukebe","url":"https://github.com/yusukebe/hono-jev-router","category":"Routing & Cost Optimization","plainSummary":"Hono 的实验性 HTTP 语义路由器，让请求描述决定走哪个处理函数。","plainSummaryEn":"An experimental semantic HTTP router for Hono using natural-language route descriptions.","jevDecisionPoint":"Jev 判断请求与各路由描述的匹配概率，程序选择首个达到阈值的路由。","jevDecisionPointEn":"Jev scores request-to-description matches; code selects the first route above its threshold.","highlightBenefit":"适合尝试按语义分流；作者明确禁止把它当作鉴权或授权边界。","highlightBenefitEn":"Explores semantic routing; the author explicitly warns against using it for authentication or authorization.","tags":["multilanguage-sdk","classification-ranking"],"stars":45,"forks":2,"openIssues":1,"license":"MIT","language":"TypeScript","lastCommitAt":"2026-09-18T08:59:00Z","createdAt":"2026-09-18T07:42:20Z","metadataFetchedAt":"2026-09-21T22:00:08.988Z","avatarUrl":"https://avatars.githubusercontent.com/u/10682?v=4","summarySource":"human-reviewed","claimStatus":"实验性语义路由；作者明确不适用于认证或授权，未验证生产流量表现。","claimStatusEn":"Experimental semantic routing, explicitly unsuitable for authentication or authorization; production traffic behavior was not verified.","plainSummaryJa":"自然言語のルート説明を使う Hono の実験的 HTTP セマンティックルーター。","plainSummaryKo":"자연어 경로 설명을 사용하는 Hono의 실험적 HTTP 의미 기반 라우터.","jevDecisionPointJa":"Jev の一致確率から、閾値を超えた最初のルートをコードが選ぶ。","jevDecisionPointKo":"Jev의 일치 확률을 이용해 임곗값을 넘는 첫 경로를 코드가 선택한다.","highlightBenefitJa":"意味による振り分けの実験用で、認証・認可の境界としての利用は禁止されている。","highlightBenefitKo":"의미 기반 분류 실험용이며 저자는 인증·인가 경계로 사용하지 말라고 명시한다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/yusukebe/hono-jev-router/blob/04f6e103e1397bca659ab85c042011a1f14b679d/README.md","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/yusukebe/hono-jev-router/blob/04f6e103e1397bca659ab85c042011a1f14b679d/src/index.ts","note":"固定版本说明与源码；本站未执行该项目。"}],"pinned":false},{"id":"razorback16:openjev","name":"openjev","author":"razorback16","url":"https://github.com/razorback16/openjev","category":"SDK & Decision Frameworks","plainSummary":"一个兼容 Jev System One 接口的独立决策服务，使用开源 DiffusionGemma 模型运行。","plainSummaryEn":"An independent System One decision server compatible with Jev’s API, running an open DiffusionGemma model.","jevDecisionPoint":"接受与 Jev 相似的状态和 Noul、Choice、Score 问题，由本地模型产生概率结果。","jevDecisionPointEn":"Accepts Jev-style state and Noul, Choice, and Score questions, producing probabilities with a local model.","highlightBenefit":"让现有 TypeSafe SDK 尝试连接自托管的兼容服务。","highlightBenefitEn":"Lets existing TypeSafe SDKs connect to a self-hosted compatible service.","tags":["multilanguage-sdk","typed-decisions"],"stars":274,"forks":23,"openIssues":2,"license":"Apache-2.0","language":"Python","lastCommitAt":"2026-09-21T16:49:03Z","createdAt":"2026-09-18T09:15:33Z","metadataFetchedAt":"2026-09-21T22:00:11.953Z","avatarUrl":"https://avatars.githubusercontent.com/u/5079550?v=4","summarySource":"human-reviewed","claimStatus":"这是协议兼容实现，不是 Jev 模型、官方产品或 TypeSafe 推理服务；速度、校准与兼容性未由本站复测。","claimStatusEn":"This is a protocol-compatible implementation, not the Jev model, an official product, or TypeSafe inference. Speed, calibration, and compatibility have not been independently retested.","plainSummaryJa":"Jev API と互換性を持ち、公開モデル DiffusionGemma で動く独立した System One サーバー。","plainSummaryKo":"Jev API와 호환되며 공개 모델 DiffusionGemma로 실행되는 독립 System One 서버.","jevDecisionPointJa":"Jev 形式の状態と質問を受け取り、ローカルモデルで確率を出す。","jevDecisionPointKo":"Jev 형식의 상태와 질문을 받고 로컬 모델로 확률을 산출한다.","highlightBenefitJa":"既存の TypeSafe SDK から自己ホストの互換サービスを試せる。","highlightBenefitKo":"기존 TypeSafe SDK로 자체 호스팅 호환 서비스를 시험할 수 있다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"confirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/razorback16/openjev/blob/91d5005effcf8cc0ecccaa9538ceabbb130fef59/openjev/api.py","note":"Fixed-commit source reviewed / 已核对固定版本源码"},{"url":"https://github.com/razorback16/openjev/blob/91d5005effcf8cc0ecccaa9538ceabbb130fef59/README.md","note":"Fixed-commit source reviewed / 已核对固定版本源码"}],"pinned":false},{"id":"giesn:typesafe-jev-workflow","name":"typesafe-jev-workflow","author":"GiesN","url":"https://github.com/GiesN/typesafe-jev-workflow","category":"Classification & Taxonomy","plainSummary":"一个异步 LangGraph 示例：让 Jev 把模拟邮件分成发票事务和普通邮件。","plainSummaryEn":"An async LangGraph example that classifies mocked emails as invoice-related or general.","jevDecisionPoint":"Jev 返回 invoice 或 general，图节点据此选择示例处理分支。","jevDecisionPointEn":"Jev returns invoice or general; graph nodes choose the corresponding demo handler.","highlightBenefit":"把模型分类和本地工作流路由分开，方便检查每步结果。","highlightBenefitEn":"Separates model classification from local workflow routing.","tags":["domain-workflows","classification-ranking"],"stars":6,"forks":1,"openIssues":0,"license":null,"language":"Python","lastCommitAt":"2026-09-16T20:54:28Z","createdAt":"2026-09-16T20:18:11Z","metadataFetchedAt":"2026-09-21T22:00:14.959Z","avatarUrl":"https://avatars.githubusercontent.com/u/61182899?v=4","summarySource":"human-reviewed","claimStatus":"输入邮件是模拟数据；handler 只设置目的地，不发送邮件或付款。","claimStatusEn":"Emails are mock data and handlers only set destinations; they do not send mail or payments. Not run here.","plainSummaryJa":"模擬メールを請求書関連と一般に分類する、非同期 LangGraph の例。","plainSummaryKo":"모의 이메일을 청구서 관련과 일반으로 분류하는 비동기 LangGraph 예제.","jevDecisionPointJa":"Jev が invoice または general を返し、graph が処理分岐を選ぶ。","jevDecisionPointKo":"Jev가 invoice 또는 general을 반환하고 graph가 처리 분기를 고른다.","highlightBenefitJa":"モデル分類とローカル workflow の振り分けを分離する。","highlightBenefitKo":"모델 분류와 로컬 workflow 라우팅을 분리한다.","claimStatusJa":"固定バージョンの文書とソースを確認。実行・性能・安全性は当サイトで独立検証していません。","claimStatusKo":"고정 버전 문서와 소스를 확인했습니다. 실행·성능·안전성은 이 사이트에서 독립적으로 검증하지 않았습니다.","licenseStatus":"unconfirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/GiesN/typesafe-jev-workflow/blob/251019670ebc3bf95870e924740d5876c1cd56b5/README.md","note":"README.md"},{"url":"https://github.com/GiesN/typesafe-jev-workflow/blob/251019670ebc3bf95870e924740d5876c1cd56b5/src/typesafe_ai_basic/__init__.py","note":"src/typesafe_ai_basic/__init__.py"},{"url":"https://github.com/GiesN/typesafe-jev-workflow/blob/251019670ebc3bf95870e924740d5876c1cd56b5/src/typesafe_ai_langgraph/typesafe_ai_langgraph_workflow.py","note":"src/typesafe_ai_langgraph/typesafe_ai_langgraph_workflow.py"}],"pinned":false},{"id":"billionsbobby:jevrouter","name":"JevRouter","author":"BillionsBobby","url":"https://github.com/BillionsBobby/JevRouter","category":"Routing & Cost Optimization","plainSummary":"把模型、Subagent、Skill、MCP 和 CLI 能力放进候选集，由 Jev 选择下一步用哪个。","plainSummaryEn":"Routes among models, subagents, skills, MCP tools and CLIs using a shared candidate set.","jevDecisionPoint":"用 Choice 评估候选能力，路由器另外检查可用性、权限、风险和确认策略。","jevDecisionPointEn":"Jev makes a Choice; the router separately checks availability, permissions, risk and confirmation policy.","highlightBenefit":"将模型选择结果与执行策略分别记录。","highlightBenefitEn":"Records model choices separately from execution policy.","tags":["llm-routing-cost","mcp-integrations","cli-git-gates"],"stars":141,"forks":11,"openIssues":6,"license":"MIT","lastCommitAt":"2026-09-21T14:48:00Z","createdAt":"2026-09-18T09:17:30Z","metadataFetchedAt":"2026-09-21T22:00:17.970Z","avatarUrl":"https://avatars.githubusercontent.com/u/133582031?v=4","summarySource":"human-reviewed","claimStatus":"已核对固定版本 README 与集成源码。","claimStatusEn":"README and integration source reviewed at a fixed commit; not independently run or benchmarked by this site.","plainSummaryJa":"モデル、Subagent、Skill、MCP、CLI を共通の候補として振り分ける。","plainSummaryKo":"모델, Subagent, Skill, MCP와 CLI를 공통 후보로 두고 라우팅한다.","jevDecisionPointJa":"Jev が Choice で選び、ルーターが可用性、権限、リスク、確認方針を別途検査する。","jevDecisionPointKo":"Jev는 Choice로 선택하고 라우터는 가용성, 권한, 위험과 확인 정책을 따로 검사한다.","highlightBenefitJa":"モデルの選択と実行方針を分けて記録する。","highlightBenefitKo":"모델 선택과 실행 정책을 분리해 기록한다.","claimStatusJa":"固定版の README と実装を確認済み。このサイトでは独立した実行・性能検証をしていません。","claimStatusKo":"고정 버전 README와 구현을 확인했습니다. 이 사이트에서 독립적인 실행·성능 검증은 하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/BillionsBobby/JevRouter/blob/715970774ae8070e958e83ac9b1a780b32a9184c/src/provider.ts#L72","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/BillionsBobby/JevRouter/blob/715970774ae8070e958e83ac9b1a780b32a9184c/README.md","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/BillionsBobby/JevRouter/blob/715970774ae8070e958e83ac9b1a780b32a9184c/docs/cookbook/README.md","note":"固定版本的实现与说明 / Pinned source and documentation"}],"pinned":false},{"id":"brainwires:jevwire","name":"jevwire","author":"Brainwires","url":"https://github.com/Brainwires/jevwire","category":"MCP & Integrations","plainSummary":"为 Agent 提供 Jev 的 MCP 工具、嵌入式库和 Claude Code hooks。","plainSummaryEn":"Provides Jev MCP tools, an embeddable library and Claude Code hooks for Agents.","jevDecisionPoint":"把排序、核验、动作检查和下一步选择交给 Jev，代码解释回答并应用策略。","jevDecisionPointEn":"Uses Jev for ranking, verification, action checks and next-step selection, with code applying policy.","highlightBenefit":"同一决策模块可从 MCP 或宿主程序调用，hooks 行为依具体配置。","highlightBenefitEn":"Reuses the decision layer through MCP or host code; hook behavior depends on configuration.","tags":["mcp-integrations","coding-agents"],"stars":15,"forks":4,"openIssues":3,"license":"MIT","language":"TypeScript","lastCommitAt":"2026-09-21T15:42:53Z","createdAt":"2026-09-17T20:56:27Z","metadataFetchedAt":"2026-09-21T22:00:20.996Z","avatarUrl":"https://avatars.githubusercontent.com/u/155631428?v=4","summarySource":"human-reviewed","claimStatus":"已核对固定版本 README 与调用源码。","claimStatusEn":"Reviewed the fixed-version README and call path.","plainSummaryJa":"Agent 向けの Jev MCP ツール、組込みライブラリ、Claude Code hooks を提供する。","plainSummaryKo":"Agent용 Jev MCP 도구·내장 라이브러리·Claude Code hooks를 제공한다.","jevDecisionPointJa":"順位付け・確認・動作点検・次の一手を Jev に問い、コードが方針を適用する。","jevDecisionPointKo":"순위·검증·동작 검사·다음 단계 선택을 Jev에 묻고 코드가 정책을 적용한다.","highlightBenefitJa":"MCP とホストコードで判断層を共用でき、hooks の動作は設定に依存する。","highlightBenefitKo":"MCP와 호스트 코드에서 판단 계층을 공유하며 hooks 동작은 설정에 따라 달라진다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/Brainwires/jevwire/blob/21abf12fabced4089779f664424fc520a67eaacb/README.md","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/Brainwires/jevwire/blob/21abf12fabced4089779f664424fc520a67eaacb/src/jev/client.ts","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/Brainwires/jevwire/blob/21abf12fabced4089779f664424fc520a67eaacb/src/tools/evaluate.ts","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/Brainwires/jevwire/blob/21abf12fabced4089779f664424fc520a67eaacb/src/hooks/handlers/pre-tool-use.ts","note":"固定版本说明与源码；本站未执行该项目。"}],"pinned":false},{"id":"wiktorb2004:llama-index-jev","name":"llama-index-jev","author":"WiktorB2004","url":"https://github.com/WiktorB2004/llama-index-jev","category":"Data & Search","plainSummary":"为 LlamaIndex 提供 Jev 重排序器和路由器，给检索片段评分或选择查询工具。","plainSummaryEn":"Provides Jev rerankers and routers for LlamaIndex to score retrieved passages or select query tools.","jevDecisionPoint":"用 Score 给候选片段评相关性，用 Choice 选择查询引擎或工具。","jevDecisionPointEn":"Uses Score for passage relevance and Choice for selecting query engines or tools.","highlightBenefit":"把 Jev 判断接入已有的检索与查询流程。","highlightBenefitEn":"Adds Jev judgments to existing retrieval and query workflows.","tags":["search-retrieval","classification-ranking","llm-routing-cost"],"stars":3,"forks":0,"openIssues":0,"license":"MIT","language":"Python","lastCommitAt":"2026-09-18T18:45:27Z","createdAt":"2026-09-18T08:19:34Z","metadataFetchedAt":"2026-09-21T22:00:23.974Z","avatarUrl":"https://avatars.githubusercontent.com/u/62223421?v=4","summarySource":"human-reviewed","claimStatus":"这是社区集成而非官方 LlamaIndex/TypeSafe 产品；重排失败时保留原序，选择器有独立失败策略。检索质量。","claimStatusEn":"A community integration, not an official LlamaIndex or TypeSafe product. Reranking preserves the retrieval order on failure; selectors have separate failure policies. Retrieval quality was not independently retested.","plainSummaryJa":"LlamaIndex 向けに、検索文章の採点と問い合わせ先の選択を行う Jev 部品を提供する。","plainSummaryKo":"LlamaIndex에서 검색 문단을 평가하고 쿼리 도구를 선택하는 Jev 구성 요소를 제공한다.","jevDecisionPointJa":"Score で文章の関連性を評価し、Choice で検索エンジンやツールを選ぶ。","jevDecisionPointKo":"Score로 문단 관련성을 평가하고 Choice로 쿼리 엔진이나 도구를 선택한다.","highlightBenefitJa":"既存の検索フローに Jev の判断を組み込める。","highlightBenefitKo":"기존 검색과 쿼리 흐름에 Jev 판단을 통합한다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"confirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/WiktorB2004/llama-index-jev/blob/72c73dc50bca4b7ea6928ef65ea09f1a7ee4a01e/packages/llama-index-postprocessor-jev/llama_index/postprocessor/jev/openrouter.py","note":"Fixed-commit source reviewed / 已核对固定版本源码"},{"url":"https://github.com/WiktorB2004/llama-index-jev/blob/72c73dc50bca4b7ea6928ef65ea09f1a7ee4a01e/README.md","note":"Fixed-commit source reviewed / 已核对固定版本源码"}],"pinned":false,"topics":["jev","llamaindex","postprocessor","python","rag","reranker","selector","typesafe"]},{"id":"abhixhek:jevcal","name":"jevcal","author":"abhixhek","url":"https://github.com/abhixhek/jevcal","category":"Evaluation & Observability","plainSummary":"用自己的标注数据评估 Jev 概率、选择置信度阈值，并检查模型更新后的变化。","plainSummaryEn":"A toolkit for evaluating Jev probabilities on labeled data, selecting confidence thresholds and checking model drift.","jevDecisionPoint":"调用决策模型处理固定问题，计算准确率、校准、覆盖率和需要升级处理的比例。","jevDecisionPointEn":"Runs fixed questions and measures accuracy, calibration, coverage and escalation rates.","highlightBenefit":"把阈值选择与模型漂移检查接到报告和 CI 中。","highlightBenefitEn":"Connects threshold selection and model-change checks to reports and CI.","tags":["evaluation-benchmarks","typed-decisions"],"stars":10,"forks":0,"openIssues":0,"license":"MIT","language":"Python","lastCommitAt":"2026-09-18T07:28:35Z","createdAt":"2026-09-18T06:50:12Z","metadataFetchedAt":"2026-09-21T22:00:26.951Z","avatarUrl":"https://avatars.githubusercontent.com/u/35265146?v=4","summarySource":"human-reviewed","claimStatus":"README 演示表使用模拟器，不是 Jev 实测成绩；生产阈值需用自己的数据验证。","claimStatusEn":"The README demo table uses a simulator, not measured Jev results; production thresholds need workload-specific validation. Not run here.","plainSummaryJa":"ラベル付きデータで Jev の確率・閾値・モデル更新の影響を調べるツール。","plainSummaryKo":"라벨 데이터로 Jev의 확률·임계값·모델 업데이트 영향을 점검하는 도구.","jevDecisionPointJa":"固定質問で精度・校正・処理範囲・追加処理率を測定する。","jevDecisionPointKo":"고정 질문으로 정확도·보정·처리 범위·상위 처리 비율을 측정한다.","highlightBenefitJa":"閾値選択とモデル変化の確認をレポートや CI に組み込める。","highlightBenefitKo":"임계값 선택과 모델 변화 확인을 보고서·CI에 연결한다.","claimStatusJa":"固定バージョンの文書とソースを確認。実行・性能・安全性は当サイトで独立検証していません。","claimStatusKo":"고정 버전 문서와 소스를 확인했습니다. 실행·성능·안전성은 이 사이트에서 독립적으로 검증하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/abhixhek/jevcal/blob/ae8f3144d69c9cb0e5e0a2c17f70b9d14714cb9f/README.md","note":"README.md"},{"url":"https://github.com/abhixhek/jevcal/blob/ae8f3144d69c9cb0e5e0a2c17f70b9d14714cb9f/src/jevcal/providers/llm.py","note":"src/jevcal/providers/llm.py"},{"url":"https://github.com/abhixhek/jevcal/blob/ae8f3144d69c9cb0e5e0a2c17f70b9d14714cb9f/src/jevcal/providers/typesafe.py","note":"src/jevcal/providers/typesafe.py"}],"pinned":false,"topics":["calibration","confidence-thresholds","jev","llm-evals","system-one","typesafe"]},{"id":"wondertwins:jev-benchmark","name":"jev-benchmark","author":"wondertwins","url":"https://github.com/wondertwins/jev-benchmark","category":"Evaluation & Observability","plainSummary":"通过国际象棋和游戏 NPC 对话对象识别，测试 Jev 的选择与判断边界。","plainSummaryEn":"Benchmarks Jev on chess moves and identifying which game NPC a player addresses.","jevDecisionPoint":"在合法棋步中选择，或分别判断玩家话语是否在对某个 NPC 说。","jevDecisionPointEn":"Selects legal chess moves or judges whether an utterance addresses each NPC.","highlightBenefit":"公开标注数据、原始请求响应与评测代码。","highlightBenefitEn":"Publishes labeled data, raw requests and responses, and evaluation code.","tags":["evaluation-benchmarks","games-simulation"],"stars":5,"forks":1,"openIssues":1,"license":"MIT","lastCommitAt":"2026-09-16T23:03:47Z","createdAt":"2026-09-16T21:12:18Z","metadataFetchedAt":"2026-09-21T22:00:30.018Z","avatarUrl":"https://avatars.githubusercontent.com/u/8965336?v=4","summarySource":"human-reviewed","claimStatus":"棋力、F1、时延均来自作者特定数据与设置，不能推广为通用游戏能力。 已核对固定版本 README 与集成源码。","claimStatusEn":"Chess strength, F1 and latency come from the author’s specific tasks and settings, not general game-playing capability. README and integration source reviewed at a fixed commit; not independently run or benchmarked by this site.","plainSummaryJa":"チェスの手選びと、話し掛けられた NPC の識別で Jev を評価する。","plainSummaryKo":"체스 수 선택과 말을 건넨 NPC 식별 과제로 Jev를 평가한다.","jevDecisionPointJa":"合法な手を選ぶか、発話が各 NPC に向けられたか判断する。","jevDecisionPointKo":"합법적인 수를 고르거나 발화가 각 NPC를 향하는지 판단한다.","highlightBenefitJa":"正解ラベル、リクエストと応答、評価コードを公開している。","highlightBenefitKo":"정답 라벨, 원본 요청·응답과 평가 코드를 공개한다.","claimStatusJa":"固定版の README と実装を確認済み。このサイトでは独立した実行・性能検証をしていません。","claimStatusKo":"고정 버전 README와 구현을 확인했습니다. 이 사이트에서 독립적인 실행·성능 검증은 하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/wondertwins/jev-benchmark/blob/1c2509ac7d5508b8a7ce00ae4df6d7652d05de8b/jevcommon/client.py#L82","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/wondertwins/jev-benchmark/blob/1c2509ac7d5508b8a7ce00ae4df6d7652d05de8b/README.md","note":"固定版本的实现与说明 / Pinned source and documentation"}],"pinned":false},{"id":"vmendes90:jev-shield","name":"jev-shield","author":"vmendes90","url":"https://github.com/vmendes90/jev-shield","category":"Browser & OS Action","plainSummary":"一个 Chrome 广告过滤扩展，用 Jev 判断信息流元素是否带有推广意图。","plainSummaryEn":"A Chrome ad-filtering extension that uses Jev to assess promotional intent in feed elements.","jevDecisionPoint":"把候选 DOM 元素分批发送到 TypeSafe，以 Noul 概率配合阈值决定是否折叠。","jevDecisionPointEn":"Batches candidate DOM elements for TypeSafe and uses Noul probabilities with thresholds to decide which to collapse.","highlightBenefit":"在本地广告规则之外增加基于文字含义的判断。","highlightBenefitEn":"Adds text-based semantic judgments alongside local ad rules.","tags":["browser-automation","classification-ranking"],"stars":2,"forks":0,"openIssues":0,"license":"MIT","language":"TypeScript","lastCommitAt":"2026-09-18T20:14:12Z","createdAt":"2026-09-18T09:12:33Z","metadataFetchedAt":"2026-09-21T22:00:33.015Z","avatarUrl":"https://avatars.githubusercontent.com/u/105750144?v=4","summarySource":"human-reviewed","claimStatus":"这是浏览器过滤工具，不是游戏或模拟项目。候选内容会发送到外部 API，误判及视频广告处理效果未由本站验证。","claimStatusEn":"This is browser filtering, not a game or simulation. Candidate content is sent to an external API.","plainSummaryJa":"情報フィードの要素が広告かを Jev で判断する Chrome 拡張。","plainSummaryKo":"피드 요소의 광고 여부를 Jev로 판단하는 Chrome 확장.","jevDecisionPointJa":"候補 DOM をまとめて TypeSafe に送り、Noul の確率と閾値で折りたたみを決める。","jevDecisionPointKo":"후보 DOM을 묶어 TypeSafe에 보내고 Noul 확률과 임계값으로 접기 여부를 결정한다.","highlightBenefitJa":"ローカルの広告ルールに意味に基づく判断を追加する。","highlightBenefitKo":"로컬 광고 규칙에 텍스트 의미 기반 판단을 추가한다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"confirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/vmendes90/jev-shield/blob/85cf3445931cbed7320f140db9af5d1fd13c4c64/src/background/typesafe.ts","note":"Fixed-commit source reviewed / 已核对固定版本源码"},{"url":"https://github.com/vmendes90/jev-shield/blob/85cf3445931cbed7320f140db9af5d1fd13c4c64/README.md","note":"Fixed-commit source reviewed / 已核对固定版本源码"}],"pinned":false,"topics":["ad-blocker","adblock","adblocker","ai","chrome-extension","manifest-v3","open-source","privacy","typesafe"]},{"id":"nainish-rai:jev-frontend-qa","name":"jev-frontend-qa","author":"Nainish-Rai","url":"https://github.com/Nainish-Rai/jev-frontend-qa","category":"Evaluation & Observability","plainSummary":"用 Jev 选择浏览器操作，再通过 DOM、HTTP 和数据库状态检查前端功能是否符合约定。","plainSummaryEn":"Frontend QA that uses Jev to choose browser actions and checks contracts through DOM, HTTP and database evidence.","jevDecisionPoint":"Jev 从已观察的控件和操作中做选择；预期值和通过条件由测试代码判断。","jevDecisionPointEn":"Jev selects observed controls and operations; test code owns expected values and pass criteria.","highlightBenefit":"将模型探索过程与可验证的功能验收分开记录。","highlightBenefitEn":"Records exploratory behavior separately from contract acceptance.","tags":["browser-automation","evaluation-benchmarks"],"stars":2,"forks":0,"openIssues":0,"license":null,"language":"Python","lastCommitAt":"2026-09-18T14:43:16Z","createdAt":"2026-09-17T20:33:24Z","metadataFetchedAt":"2026-09-21T22:00:36.011Z","avatarUrl":"https://avatars.githubusercontent.com/u/109546113?v=4","summarySource":"human-reviewed","claimStatus":"作者报告的是合成参考场景验收，不等于任意网站都通过；本站未重跑。","claimStatusEn":"The author reports acceptance on synthetic reference scenarios, not arbitrary websites; not rerun here.","plainSummaryJa":"Jev がブラウザー操作を選び、DOM・HTTP・DB の証拠で仕様を確認するフロントエンド QA。","plainSummaryKo":"Jev가 브라우저 동작을 고르고 DOM·HTTP·DB 증거로 계약을 확인하는 프런트엔드 QA.","jevDecisionPointJa":"観測した操作と操作部品から選び、期待値と合格条件はテストコードが判定する。","jevDecisionPointKo":"관측한 컨트롤과 동작을 선택하며 기대값과 통과 기준은 테스트 코드가 판단한다.","highlightBenefitJa":"探索の記録と仕様の受入確認を区別できる。","highlightBenefitKo":"탐색 기록과 계약 수락 검사를 구분한다.","claimStatusJa":"固定バージョンの文書とソースを確認。実行・性能・安全性は当サイトで独立検証していません。","claimStatusKo":"고정 버전 문서와 소스를 확인했습니다. 실행·성능·안전성은 이 사이트에서 독립적으로 검증하지 않았습니다.","licenseStatus":"unconfirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/Nainish-Rai/jev-frontend-qa/blob/2a57d0a23471ca627a2840eff04102ddc844dd57/README.md","note":"README.md"},{"url":"https://github.com/Nainish-Rai/jev-frontend-qa/blob/2a57d0a23471ca627a2840eff04102ddc844dd57/src/jev_frontend_qa/core/decisions.py","note":"src/jev_frontend_qa/core/decisions.py"},{"url":"https://github.com/Nainish-Rai/jev-frontend-qa/blob/2a57d0a23471ca627a2840eff04102ddc844dd57/src/jev_frontend_qa/core/model_client.py","note":"src/jev_frontend_qa/core/model_client.py"}],"pinned":false},{"id":"jerryfane:omp-jev-compaction","name":"omp-jev-compaction","author":"jerryfane","url":"https://github.com/jerryfane/omp-jev-compaction","category":"Context GC & Filter","plainSummary":"为 Oh My Pi 删减工具历史的扩展，保留原文并复用旧判断以减少前缀反复改写。","plainSummaryEn":"An Oh My Pi extension that prunes tool history verbatim and reuses decisions to limit prefix rewrites.","jevDecisionPoint":"Jev 判断工具调用和结果的保留价值，扩展据此截短并附上可恢复提示。","jevDecisionPointEn":"Jev judges tool-call and output utility; the extension truncates selected content with recovery notes.","highlightBenefit":"将删减决策记下来，在后续请求中重复应用。","highlightBenefitEn":"Remembers pruning decisions and reapplies them across requests.","tags":["context-compaction","coding-agents"],"stars":8,"forks":1,"openIssues":3,"license":"MIT","lastCommitAt":"2026-09-19T08:53:47Z","createdAt":"2026-09-18T08:09:25Z","metadataFetchedAt":"2026-09-21T22:00:39.000Z","avatarUrl":"https://avatars.githubusercontent.com/u/38183696?v=4","summarySource":"human-reviewed","claimStatus":"已核对固定版本 README 与集成源码。","claimStatusEn":"README and integration source reviewed at a fixed commit; not independently run or benchmarked by this site.","plainSummaryJa":"Oh My Pi のツール履歴を整理し、判断を再利用して先頭部分の書き換えを抑える拡張。","plainSummaryKo":"Oh My Pi 도구 기록을 정리하고 판단을 재사용해 문맥 앞부분의 반복 수정을 줄이는 확장이다.","jevDecisionPointJa":"Jev が呼び出しと結果の必要性を判断し、選択した内容を復元用の注記付きで短縮する。","jevDecisionPointKo":"Jev가 호출과 결과의 필요성을 판단하면 선택된 내용을 복원 안내와 함께 축약한다.","highlightBenefitJa":"削減判断を記憶して後のリクエストに適用する。","highlightBenefitKo":"정리 판단을 기억해 이후 요청에 다시 적용한다.","claimStatusJa":"固定版の README と実装を確認済み。このサイトでは独立した実行・性能検証をしていません。","claimStatusKo":"고정 버전 README와 구현을 확인했습니다. 이 사이트에서 독립적인 실행·성능 검증은 하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/jerryfane/omp-jev-compaction/blob/5cc9daf51d97f0922d39f8396694631d6a87535c/src/asker.ts#L15","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/jerryfane/omp-jev-compaction/blob/5cc9daf51d97f0922d39f8396694631d6a87535c/README.md","note":"固定版本的实现与说明 / Pinned source and documentation"}],"pinned":false},{"id":"maker-kk:todo-jev","name":"todo-jev","author":"maker-KK","url":"https://github.com/maker-KK/todo-jev","category":"Routing & Cost Optimization","plainSummary":"结合 skill 条件和环境检查的任务路由实验，推荐规则、skill 或大模型处理路径。","plainSummaryEn":"A task-routing experiment combining skill conditions and environment checks to suggest rules, skills or a large model.","jevDecisionPoint":"Jev 分类请求并匹配候选 skill；没有密钥或调用失败时使用启发式回退。","jevDecisionPointEn":"Jev classifies requests and matches skills, with heuristic fallback when unavailable.","highlightBenefit":"已实现分类与推荐，执行处理器仍返回示例响应，需要自行接入实际执行。","highlightBenefitEn":"Classification and recommendations are implemented; execution handlers still return examples and need real integrations.","tags":["llm-routing-cost","typed-decisions"],"stars":2,"forks":1,"openIssues":1,"license":"MIT","language":"Python","lastCommitAt":"2026-09-18T11:41:03Z","createdAt":"2026-09-18T08:52:14Z","metadataFetchedAt":"2026-09-21T22:00:41.960Z","avatarUrl":"https://avatars.githubusercontent.com/u/263155563?v=4","summarySource":"human-reviewed","claimStatus":"已核对真实分类调用；任务执行处理器仍为示例输出，未验证完整任务执行。","claimStatusEn":"The classification call is implemented, but task execution handlers return example output; end-to-end task execution is unverified.","plainSummaryJa":"skill の条件と環境確認から、ルール・skill・大規模モデルを勧めるタスクルーティング実験。","plainSummaryKo":"skill 조건과 환경 점검으로 규칙·skill·대형 모델 경로를 추천하는 작업 라우팅 실험.","jevDecisionPointJa":"Jev が要求を分類して skill を照合し、利用不可時はヒューリスティックへ戻る。","jevDecisionPointKo":"Jev가 요청을 분류하고 skill을 매칭하며 사용 불가 시 휴리스틱으로 대체한다.","highlightBenefitJa":"分類と推薦は実装済みだが、実行ハンドラーは例示応答で実処理の接続が必要。","highlightBenefitKo":"분류·추천은 구현됐으나 실행 핸들러는 예제 응답이므로 실제 실행 연결이 필요하다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/maker-KK/todo-jev/blob/08c8a1e744c311e9f88721b414e636d7eb7c156c/README.md","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/maker-KK/todo-jev/blob/08c8a1e744c311e9f88721b414e636d7eb7c156c/app/classifier.py","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/maker-KK/todo-jev/blob/08c8a1e744c311e9f88721b414e636d7eb7c156c/app/router.py","note":"固定版本说明与源码；本站未执行该项目。"}],"pinned":false},{"id":"ozers:jevsome-projects","name":"jevsome-projects","author":"ozers","url":"https://github.com/ozers/jevsome-projects","category":"Data & Search","plainSummary":"一个 Jev 项目目录与发现流水线，保存接入证据，并可用 Jev 辅助分类。","plainSummaryEn":"A Jev project directory and discovery pipeline that stores integration evidence and can use Jev for classification.","jevDecisionPoint":"配置 key 后，把项目状态和候选类别交给 Jev；未配置时使用本地分类规则。","jevDecisionPointEn":"When configured with a key, sends project state and candidate categories to Jev; otherwise uses local classification rules.","highlightBenefit":"把项目索引与具体源码证据放在一起。","highlightBenefitEn":"Keeps a project index alongside concrete source evidence.","tags":["search-retrieval","classification-ranking"],"stars":1,"forks":1,"openIssues":1,"license":"MIT","language":"JavaScript","lastCommitAt":"2026-09-18T15:14:11Z","createdAt":"2026-09-18T11:15:13Z","metadataFetchedAt":"2026-09-21T22:00:45.013Z","avatarUrl":"https://avatars.githubusercontent.com/u/22887208?v=4","summarySource":"human-reviewed","claimStatus":"存在实际 TypeSafe 调用代码，但目录对下游项目的评级与证据仍由该目录作者负责，本站未逐项认可其结论。","claimStatusEn":"Actual TypeSafe request code exists, but the directory’s ratings and evidence for downstream projects remain its authors’ assessments; this site does not endorse each conclusion.","plainSummaryJa":"接続の根拠を保存し、任意で Jev に分類させるプロジェクト一覧と探索パイプライン。","plainSummaryKo":"연동 근거를 저장하고 선택적으로 Jev로 분류하는 프로젝트 목록과 탐색 파이프라인.","jevDecisionPointJa":"キー設定時は状態と分類候補を Jev に送り、未設定時はローカルルールを使う。","jevDecisionPointKo":"키 설정 시 상태와 범주 후보를 Jev에 보내고 미설정 시 로컬 규칙을 사용한다.","highlightBenefitJa":"プロジェクト一覧と具体的なコード上の根拠をまとめる。","highlightBenefitKo":"프로젝트 목록과 구체적인 소스 근거를 함께 제공한다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"confirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/ozers/jevsome-projects/blob/ba0bfdceb27f8ba1df10f76d4b5b090438eee952/pipeline/lib/jev.mjs","note":"Fixed-commit source reviewed / 已核对固定版本源码"},{"url":"https://github.com/ozers/jevsome-projects/blob/ba0bfdceb27f8ba1df10f76d4b5b090438eee952/README.md","note":"Fixed-commit source reviewed / 已核对固定版本源码"}],"pinned":false,"topics":["ai","awesome","awesome-list","directory","jev","system-one","typed-decisions","typesafe-ai"]},{"id":"nasrallah-al:jev-cli","name":"jev-cli","author":"Nasrallah-AL","url":"https://github.com/Nasrallah-AL/jev-cli","category":"CLI & Pipelines","plainSummary":"名为 jevctl 的终端工具，把核验、分类与评分问题接到文本输入和脚本里。","plainSummaryEn":"The jevctl command-line tool connects verification, classification and scoring questions to text input and scripts.","jevDecisionPoint":"把输入与固定答案集合发给 Jev，再按本地阈值输出判断和概率。","jevDecisionPointEn":"Sends input and bounded answer options to Jev, then reports judgments and probabilities using local thresholds.","highlightBenefit":"提供可用于管道和 CI 的结构化结果，也支持检查请求和 dry-run。","highlightBenefitEn":"Provides structured results for pipelines and CI, with request inspection and dry-run support.","tags":["cli-git-gates","classification-ranking","typed-decisions"],"stars":18,"forks":1,"openIssues":0,"license":"MIT","language":"TypeScript","lastCommitAt":"2026-09-20T20:25:13Z","createdAt":"2026-09-18T08:16:41Z","metadataFetchedAt":"2026-09-21T22:00:47.934Z","avatarUrl":"https://avatars.githubusercontent.com/u/114925308?v=4","summarySource":"human-reviewed","claimStatus":"与其他作者的同名 jev-cli 是不同仓库；仅核对实现，未测性能或核验准确率。","claimStatusEn":"Distinct from other authors’ jev-cli repositories. Implementation was inspected; performance and verification accuracy were not measured.","plainSummaryJa":"検証・分類・評価の質問を文章入力やスクリプトへ接続する jevctl CLI。","plainSummaryKo":"검증·분류·평가 질문을 텍스트 입력과 스크립트에 연결하는 jevctl CLI.","jevDecisionPointJa":"入力と固定候補を Jev に送り、ローカル閾値で判断と確率を返す。","jevDecisionPointKo":"입력과 고정 후보를 Jev에 보내고 로컬 임계값으로 판단과 확률을 반환한다.","highlightBenefitJa":"パイプライン・CI 用の結果と、リクエスト確認・dry-run を提供する。","highlightBenefitKo":"파이프라인·CI용 결과와 요청 확인·dry-run을 제공한다.","claimStatusJa":"固定バージョンの文書とソースを確認。実行・性能・安全性は当サイトで独立検証していません。","claimStatusKo":"고정 버전 문서와 소스를 확인했습니다. 실행·성능·안전성은 이 사이트에서 독립적으로 검증하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/Nasrallah-AL/jev-cli/blob/02ca80177aaef0b30a897959207f99df1e0c8446/README.md","note":"README.md"},{"url":"https://github.com/Nasrallah-AL/jev-cli/blob/02ca80177aaef0b30a897959207f99df1e0c8446/plugin/hooks/fast-jev.ts","note":"plugin/hooks/fast-jev.ts"},{"url":"https://github.com/Nasrallah-AL/jev-cli/blob/02ca80177aaef0b30a897959207f99df1e0c8446/src/provider.ts","note":"src/provider.ts"}],"pinned":false,"topics":["ai","cli","fact-checking","guardrails","jev","llm-tools","nodejs","prompt-injection","semantic-search","typesafe","typescript"]},{"id":"poponline63:hermes-jev-north-star","name":"hermes-jev-north-star","author":"poponline63","url":"https://github.com/poponline63/hermes-jev-north-star","category":"Evaluation & Observability","plainSummary":"Hermes 的目标验收 skill，保存要求、生成运行提示词，并检查证据是否满足要求。","plainSummaryEn":"A Hermes goal-checking skill that saves requirements, creates a run prompt and checks completion evidence.","jevDecisionPoint":"可机器检查的部分运行本地检查，Jev 对其余语义要求判断达成程度。","jevDecisionPointEn":"Local checks handle machine-verifiable requirements; Jev assesses the remaining semantic conditions.","highlightBenefit":"把目标与可检查条件放在一起；模型判断不替代真实验收证据。","highlightBenefitEn":"Connects goals with checkable criteria without replacing actual acceptance evidence with model opinions.","tags":["evaluation-benchmarks","mcp-integrations"],"stars":1,"forks":1,"openIssues":0,"license":"MIT","language":"Python","lastCommitAt":"2026-09-18T18:41:35Z","createdAt":"2026-09-18T10:37:38Z","metadataFetchedAt":"2026-09-21T22:00:50.980Z","avatarUrl":"https://avatars.githubusercontent.com/u/232531104?v=4","summarySource":"human-reviewed","claimStatus":"已核对固定版本 README 与调用源码。","claimStatusEn":"Reviewed the fixed-version README and call path.","plainSummaryJa":"要件を保存し実行プロンプトを作り、完了証拠を点検する Hermes の目標確認 skill。","plainSummaryKo":"요구사항을 저장하고 실행 프롬프트를 만들며 완료 증거를 점검하는 Hermes skill.","jevDecisionPointJa":"機械で確認できる要件はローカルで、残りの意味条件は Jev で評価する。","jevDecisionPointKo":"기계 검증 가능한 요구는 로컬에서, 나머지 의미 조건은 Jev가 평가한다.","highlightBenefitJa":"目標を検査条件に結び付けるが、実際の受入証拠をモデル意見で置き換えない。","highlightBenefitKo":"목표를 검증 조건에 연결하되 실제 인수 증거를 모델 의견으로 대체하지 않는다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/poponline63/hermes-jev-north-star/blob/b5499f2da6b940258e2dc7c935552a217c7df993/README.md","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/poponline63/hermes-jev-north-star/blob/b5499f2da6b940258e2dc7c935552a217c7df993/scripts/jev_judge.py","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/poponline63/hermes-jev-north-star/blob/b5499f2da6b940258e2dc7c935552a217c7df993/scripts/north_star.py","note":"固定版本说明与源码；本站未执行该项目。"}],"pinned":false,"topics":["agent-skills","ai-agents","autonomy","definition-of-done","hermes-agent","jev","system-one","typesafe"]},{"id":"wang-auspicious:codex-jev-compaction","name":"codex-jev-compaction","author":"Wang-auspicious","url":"https://github.com/Wang-auspicious/codex-jev-compaction","category":"Context GC & Filter","plainSummary":"为 Codex 整理任务交接上下文，用 Jev 筛选旧工具记录，并保留选中内容的原文。","plainSummaryEn":"Curates Codex handoff context by using Jev to select old tool records while retaining selected text verbatim.","jevDecisionPoint":"判断符合条件的只读工具记录是否相关，本地规则保护必须保留的内容并生成交接包。","jevDecisionPointEn":"Judges relevance of eligible read-only tool records; local rules protect required content and assemble a handoff packet.","highlightBenefit":"交接包保留来源、筛选理由和原文次序。","highlightBenefitEn":"Preserves source references, selection reasons, and original text order in the handoff.","tags":["context-compaction","coding-agents","mcp-integrations"],"stars":2,"forks":0,"openIssues":0,"license":"MIT","language":"JavaScript","lastCommitAt":"2026-09-18T11:08:29Z","createdAt":"2026-09-18T10:43:59Z","metadataFetchedAt":"2026-09-21T22:00:53.961Z","avatarUrl":"https://avatars.githubusercontent.com/u/198441379?v=4","summarySource":"human-reviewed","claimStatus":"准备显式交接包，不替代 Codex 原生自动压缩，也不回收当前会话 Token；筛选有损，原始输入仍应保留。","claimStatusEn":"Prepares explicit handoff packets; it does not replace Codex automatic compaction or reclaim Tokens in the current conversation. Selection is lossy, so original input should be retained.","plainSummaryJa":"Jev で過去のツール記録を選別し、採用した原文を残す Codex 向け引き継ぎ支援。","plainSummaryKo":"Jev로 이전 도구 기록을 선별하고 선택한 원문을 유지하는 Codex 인계 도구.","jevDecisionPointJa":"対象となる読み取り専用記録の関連性を判断し、必須内容を保護して引き継ぎ資料を作る。","jevDecisionPointKo":"대상 읽기 전용 기록의 관련성을 판단하고 필수 내용을 보호해 인계 자료를 만든다.","highlightBenefitJa":"出典、選別理由、原文の順序を残す。","highlightBenefitKo":"출처, 선별 이유, 원문 순서를 보존한다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"confirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/Wang-auspicious/codex-jev-compaction/blob/26fa09f83113a777b0d0db25d847f1c2a25d3388/plugins/codex-jev-compaction/lib/core.mjs","note":"Fixed-commit source reviewed / 已核对固定版本源码"},{"url":"https://github.com/Wang-auspicious/codex-jev-compaction/blob/26fa09f83113a777b0d0db25d847f1c2a25d3388/README.md","note":"Fixed-commit source reviewed / 已核对固定版本源码"}],"pinned":false,"topics":["ai-agents","codex","codex-plugin","context-compaction","jev","mcp","typesafe-ai"]},{"id":"wang-auspicious:pi-jev-compaction","name":"pi-jev-compaction","author":"Wang-auspicious","url":"https://github.com/Wang-auspicious/pi-jev-compaction","category":"Context GC & Filter","plainSummary":"给 Pi 做抽取式上下文压缩：让 Jev 判断旧的只读工具记录是否仍有用，保留部分直接复制原文。","plainSummaryEn":"Extractive context compaction for Pi that keeps selected original tool records instead of generating a summary.","jevDecisionPoint":"以完整工具调用与返回为单位判断相关性，由代码移除被判为无用的记录对。","jevDecisionPointEn":"Jev judges complete read-only tool pairs; local code removes those rated unnecessary.","highlightBenefit":"保留原文证据和 Pi 的近期消息边界，不让 Jev 改写摘要。","highlightBenefitEn":"Preserves original evidence and Pi’s recent-message boundary.","tags":["context-compaction","coding-agents"],"stars":1,"forks":0,"openIssues":0,"license":"MIT","language":"TypeScript","lastCommitAt":"2026-09-18T10:59:31Z","createdAt":"2026-09-18T10:43:18Z","metadataFetchedAt":"2026-09-21T22:00:56.970Z","avatarUrl":"https://avatars.githubusercontent.com/u/198441379?v=4","summarySource":"human-reviewed","claimStatus":"依赖 Pi 的 compaction 接口及特定版本；抽取筛选仍可能遗漏有用内容。","claimStatusEn":"Depends on Pi’s compaction interface and version; extractive filtering can still omit useful content. Not run here.","plainSummaryJa":"生成要約ではなく、選んだツール記録の原文を残す Pi 向け Context GC。","plainSummaryKo":"요약을 생성하는 대신 선택한 도구 기록의 원문을 남기는 Pi용 Context GC.","jevDecisionPointJa":"読み取り専用ツールの呼び出しと返答を一組で判定し、不要な組をコードで除く。","jevDecisionPointKo":"읽기 전용 도구 호출과 결과를 한 쌍으로 판단하고 불필요한 쌍을 코드가 제거한다.","highlightBenefitJa":"原文の証拠と Pi の最近のメッセージ境界を残す。","highlightBenefitKo":"원문 증거와 Pi의 최근 메시지 경계를 보존한다.","claimStatusJa":"固定バージョンの文書とソースを確認。実行・性能・安全性は当サイトで独立検証していません。","claimStatusKo":"고정 버전 문서와 소스를 확인했습니다. 실행·성능·안전성은 이 사이트에서 독립적으로 검증하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/Wang-auspicious/pi-jev-compaction/blob/058d91db752f97c7d52062e465eb151bddab060d/README.md","note":"README.md"},{"url":"https://github.com/Wang-auspicious/pi-jev-compaction/blob/058d91db752f97c7d52062e465eb151bddab060d/src/client.ts","note":"src/client.ts"}],"pinned":false,"topics":["ai-agents","context-compaction","jev","pi","pi-package","typesafe-ai","typescript"]},{"id":"fatwang2:jev-review-action","name":"jev-review-action","author":"fatwang2","url":"https://github.com/fatwang2/jev-review-action","category":"Codebase & Graph Pathfinding","plainSummary":"可配置的 GitHub Action，用 Jev 检查目录投稿或给 PR 分类，并更新模板评论。","plainSummaryEn":"A configurable GitHub Action that reviews catalog submissions or classifies PRs and updates a templated comment.","jevDecisionPoint":"依据固定版本仓库材料或 PR patch 回答策略问题，再由代码应用分类规则。","jevDecisionPointEn":"Answers policy questions over pinned repository evidence or PR patches, then applies classification rules.","highlightBenefit":"把问题定义、阈值与评论格式留在可审查的配置中。","highlightBenefitEn":"Keeps question definitions, thresholds and comment templates inspectable.","tags":["cli-git-gates","coding-agents","classification-ranking"],"stars":1,"forks":1,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-20T19:04:50Z","createdAt":"2026-09-18T02:44:49Z","metadataFetchedAt":"2026-09-21T22:01:00.096Z","avatarUrl":"https://avatars.githubusercontent.com/u/134143178?v=4","summarySource":"human-reviewed","claimStatus":"源码支持审查、分类与评论，不等于自动合并 PR；旧英文 auto-merge 说法已删除。 已核对固定版本 README 与集成源码。","claimStatusEn":"The code reviews, classifies and comments; it does not establish automatic PR merging. The previous auto-merge description was removed. README and integration source reviewed at a fixed commit; not independently run or benchmarked by this site.","plainSummaryJa":"ディレクトリ投稿の確認や PR 分類を行い、定型コメントを更新する GitHub Action。","plainSummaryKo":"목록 제출을 검토하거나 PR을 분류하고 서식 댓글을 갱신하는 GitHub Action이다.","jevDecisionPointJa":"固定版の根拠や PR 差分について方針の質問に答え、コードが分類規則を適用する。","jevDecisionPointKo":"고정 버전 근거나 PR diff로 정책 질문에 답하고 코드가 분류 규칙을 적용한다.","highlightBenefitJa":"質問、しきい値、コメント形式を確認可能な設定に置く。","highlightBenefitKo":"질문, 임계값과 댓글 서식을 검토 가능한 설정에 둔다.","claimStatusJa":"固定版の README と実装を確認済み。このサイトでは独立した実行・性能検証をしていません。","claimStatusKo":"고정 버전 README와 구현을 확인했습니다. 이 사이트에서 독립적인 실행·성능 검증은 하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/fatwang2/jev-review-action/blob/0cc579cb2aa5b160b6d8a788a4775953a31c6e1e/src/typesafe.mjs#L36","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/fatwang2/jev-review-action/blob/0cc579cb2aa5b160b6d8a788a4775953a31c6e1e/README.md","note":"固定版本的实现与说明 / Pinned source and documentation"}],"pinned":false},{"id":"akashpriyadarshii:jev-curate","name":"jev-curate","author":"AkashPriyadarshii","url":"https://github.com/AkashPriyadarshii/jev-curate","category":"Data & Search","plainSummary":"Rust 数据集筛选实验，用 Jev 给文本记录评分并分流到保留或拒绝结果。","plainSummaryEn":"A Rust dataset-filtering experiment using Jev scores to keep or reject text records.","jevDecisionPoint":"本地预过滤后调用 TypeSafe，按各项概率和评分阈值决定是否保留记录。","jevDecisionPointEn":"Runs local prefilters, calls TypeSafe and applies probability and score thresholds to each record.","highlightBenefit":"提供本地预过滤与分数阈值，适合研究按记录筛选的数据管道。","highlightBenefitEn":"Provides local prefilters and score thresholds for experimenting with per-record data pipelines.","tags":["classification-ranking","evaluation-benchmarks"],"stars":16,"forks":1,"openIssues":1,"license":"MIT","language":"Rust","lastCommitAt":"2026-09-21T17:52:41Z","createdAt":"2026-09-18T06:36:41Z","metadataFetchedAt":"2026-09-21T22:01:03.069Z","avatarUrl":"https://avatars.githubusercontent.com/u/272530059?v=4","summarySource":"human-reviewed","claimStatus":"已核对 Rust API 调用与筛选路径；README 的 1500+ rows/sec 无可复核测量，Python sift 方法未在固定版绑定实现。","claimStatusEn":"Reviewed the Rust API and filtering path; the advertised 1,500+ rows/sec has no verified measurement and the checked Python binding lacks the documented sift method.","plainSummaryJa":"Jev の採点でテキストレコードを採用・除外する Rust のデータセット選別実験。","plainSummaryKo":"Jev 점수로 텍스트 레코드를 유지·제외하는 Rust 데이터셋 필터링 실험.","jevDecisionPointJa":"ローカル前処理後に TypeSafe を呼び、確率とスコアの閾値を適用する。","jevDecisionPointKo":"로컬 사전 필터 후 TypeSafe를 호출하고 확률·점수 임곗값을 적용한다.","highlightBenefitJa":"ローカル前処理とスコア閾値を備え、レコード単位の選別パイプラインを研究できる。","highlightBenefitKo":"로컬 사전 필터와 점수 임곗값으로 레코드 단위 선별 파이프라인을 실험할 수 있다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/AkashPriyadarshii/jev-curate/blob/a35d420cdf15eb1d109171fc896ddc54e6cb9a45/README.md","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/AkashPriyadarshii/jev-curate/blob/a35d420cdf15eb1d109171fc896ddc54e6cb9a45/src/client.rs","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/AkashPriyadarshii/jev-curate/blob/a35d420cdf15eb1d109171fc896ddc54e6cb9a45/src/filter.rs","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/AkashPriyadarshii/jev-curate/blob/a35d420cdf15eb1d109171fc896ddc54e6cb9a45/src/main.rs","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/AkashPriyadarshii/jev-curate/blob/a35d420cdf15eb1d109171fc896ddc54e6cb9a45/src/lib.rs","note":"固定版本说明与源码；本站未执行该项目。"}],"pinned":false,"topics":["arrow","cli","data-cleaning","data-engineering","dataset-curation","eval-harness","fine-tuning","jev","jsonl","llm","parquet","polars","pretraining","pyo3","reasoning-models","rust","synthetic-data","system-one","type-safe-ai","typesafe-ai"]},{"id":"altryne:jevify","name":"jevify","author":"altryne","url":"https://github.com/altryne/jevify","category":"Decision Tools","plainSummary":"一个 Agent Skill，帮助找出项目中适合用 Jev 的判断环节，并设计问题与对照实验。","plainSummaryEn":"An Agent Skill for finding suitable Jev decision points and designing questions and comparison experiments.","jevDecisionPoint":"围绕应用场景设计 Noul、Choice、Score 问题包，附脚本可调用 API 运行案例。","jevDecisionPointEn":"Designs Noul, Choice, and Score question packs for an application, with a script for API-backed case runs.","highlightBenefit":"把接入想法、问题设计与验证方法放在一起。","highlightBenefitEn":"Connects integration ideas, question design, and evaluation methods.","tags":["mcp-integrations","evaluation-benchmarks","typed-decisions"],"stars":21,"forks":2,"openIssues":1,"license":"MIT","language":"Python","lastCommitAt":"2026-09-21T08:00:29Z","createdAt":"2026-09-17T19:54:57Z","metadataFetchedAt":"2026-09-21T22:01:05.946Z","avatarUrl":"https://avatars.githubusercontent.com/u/463317?v=4","summarySource":"human-reviewed","claimStatus":"不是 Jev 模型、SDK 或官方产品；安装 Skill 本身不会调用 API，真实推理需另行执行并配置 key。","claimStatusEn":"Not the Jev model, an SDK, or an official product. Installing the Skill makes no API calls; actual inference requires a separate run and configured key.","plainSummaryJa":"Jev に適した判断箇所を探し、質問と比較実験を設計する Agent Skill。","plainSummaryKo":"Jev에 적합한 판단 지점을 찾고 질문과 비교 실험을 설계하는 Agent Skill.","jevDecisionPointJa":"用途に応じた質問を作り、付属スクリプトで API を使うケースを実行できる。","jevDecisionPointKo":"용도에 맞는 질문을 만들며 포함된 스크립트로 API 기반 사례를 실행할 수 있다.","highlightBenefitJa":"導入案、質問設計、評価方法を結び付ける。","highlightBenefitKo":"연동 아이디어, 질문 설계, 평가 방법을 연결한다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"confirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/altryne/jevify/blob/9b50ba134487f9a5a568386583c18e876cae3f56/scripts/run_cases.py","note":"Fixed-commit source reviewed / 已核对固定版本源码"},{"url":"https://github.com/altryne/jevify/blob/9b50ba134487f9a5a568386583c18e876cae3f56/README.md","note":"Fixed-commit source reviewed / 已核对固定版本源码"}],"pinned":false,"topics":["agent-skills","ai-agents","jev","thursdai","typesafe"]},{"id":"pistachioaihq:jev-synergy-screening","name":"jev-synergy-screening","author":"PistachioAIHQ","url":"https://github.com/PistachioAIHQ/jev-synergy-screening","category":"Evaluation & Observability","plainSummary":"用 Jev 筛选 ADHD 综述的论文标题和摘要，并与 Cohen Abstract Triage 标注比较。","plainSummaryEn":"A Jev title-and-abstract screening experiment compared with Cohen Abstract Triage labels for an ADHD review.","jevDecisionPoint":"围绕纳入标准回答 Choice 与 Noul 问题，本地规则组合成 include 或 exclude。","jevDecisionPointEn":"Asks Choice and Noul questions about eligibility, then combines them into include or exclude decisions.","highlightBenefit":"保留不同数据切片和问法的指标，便于检查漏筛与误筛。","highlightBenefitEn":"Records metrics for specific dataset slices and question designs to examine screening errors.","tags":["domain-workflows","search-retrieval","evaluation-benchmarks"],"stars":1,"forks":1,"openIssues":0,"license":null,"language":"Python","lastCommitAt":"2026-09-16T22:52:02Z","createdAt":"2026-09-16T20:44:13Z","metadataFetchedAt":"2026-09-21T22:01:08.991Z","avatarUrl":"https://avatars.githubusercontent.com/u/250697747?v=4","summarySource":"human-reviewed","claimStatus":"特定文献数据集实验，不能视作临床结论或替代系统综述人工审核；本站未重跑。","claimStatusEn":"A dataset-specific literature experiment, not clinical evidence or a replacement for human systematic-review screening; not rerun here.","plainSummaryJa":"ADHD レビューの題名・抄録を Jev で選別し、Cohen Abstract Triage のラベルと比較する実験。","plainSummaryKo":"ADHD 리뷰의 제목·초록을 Jev로 선별하고 Cohen Abstract Triage 라벨과 비교하는 실험.","jevDecisionPointJa":"適格性を Choice と Noul で質問し、コードで採用・除外にまとめる。","jevDecisionPointKo":"적격성을 Choice와 Noul로 묻고 코드가 포함·제외로 조합한다.","highlightBenefitJa":"データ区分と質問設計ごとの指標から選別ミスを確認できる。","highlightBenefitKo":"데이터 구간과 질문 설계별 지표로 선별 오류를 확인한다.","claimStatusJa":"固定バージョンの文書とソースを確認。実行・性能・安全性は当サイトで独立検証していません。","claimStatusKo":"고정 버전 문서와 소스를 확인했습니다. 실행·성능·안전성은 이 사이트에서 독립적으로 검증하지 않았습니다.","licenseStatus":"unconfirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/PistachioAIHQ/jev-synergy-screening/blob/4142c88513b68520958cee1fa72ed59385f91376/README.md","note":"README.md"},{"url":"https://github.com/PistachioAIHQ/jev-synergy-screening/blob/4142c88513b68520958cee1fa72ed59385f91376/src/jev_client.py","note":"src/jev_client.py"}],"pinned":false},{"id":"byk:jev-mcp","name":"jev-mcp","author":"BYK","url":"https://github.com/BYK/jev-mcp","category":"MCP & Integrations","plainSummary":"以评测为重点的 Jev MCP 服务，可单次提问、批量处理并比较问题和阈值。","plainSummaryEn":"An evaluation-focused Jev MCP server for individual questions, batch processing and question or threshold comparisons.","jevDecisionPoint":"对状态运行 Choice、Score、Noul，再用标注样本计算准确率与校准等指标。","jevDecisionPointEn":"Runs typed questions and uses labeled samples to measure accuracy and calibration.","highlightBenefit":"让问题设计与阈值选择有可检查的样本结果。","highlightBenefitEn":"Makes question design and threshold choices inspectable through sample results.","tags":["mcp-integrations","evaluation-benchmarks","typed-decisions"],"stars":1,"forks":0,"openIssues":1,"license":"MIT","lastCommitAt":"2026-09-18T09:14:03Z","createdAt":"2026-09-17T21:59:18Z","metadataFetchedAt":"2026-09-21T22:01:12.019Z","avatarUrl":"https://avatars.githubusercontent.com/u/126780?v=4","summarySource":"human-reviewed","claimStatus":"已核对固定版本 README 与集成源码。","claimStatusEn":"README and integration source reviewed at a fixed commit; not independently run or benchmarked by this site.","plainSummaryJa":"単発質問、一括処理、質問やしきい値の比較を行う評価重視の Jev MCP サーバー。","plainSummaryKo":"개별 질문, 일괄 처리, 질문·임계값 비교를 제공하는 평가 중심 Jev MCP 서버다.","jevDecisionPointJa":"型付き質問を実行し、正解付き標本で正確さや較正を測る。","jevDecisionPointKo":"구조화된 질문을 실행하고 정답 표본으로 정확도와 보정을 측정한다.","highlightBenefitJa":"標本結果を使って質問設計としきい値を確認できる。","highlightBenefitKo":"표본 결과로 질문 설계와 임계값을 검토할 수 있다.","claimStatusJa":"固定版の README と実装を確認済み。このサイトでは独立した実行・性能検証をしていません。","claimStatusKo":"고정 버전 README와 구현을 확인했습니다. 이 사이트에서 독립적인 실행·성능 검증은 하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/BYK/jev-mcp/blob/cee2e6d58112d006ccd3ebad1163a31083262ae9/src/typesafe.ts#L1","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/BYK/jev-mcp/blob/cee2e6d58112d006ccd3ebad1163a31083262ae9/README.md","note":"固定版本的实现与说明 / Pinned source and documentation"}],"pinned":false},{"id":"felpsdev:jev-classifier","name":"jev-classifier","author":"felpsdev","url":"https://github.com/felpsdev/jev-classifier","category":"MCP & Integrations","plainSummary":"连接编码 Agent 的本地 Jev 工具路由网关，同时提供 MCP 建议接口。","plainSummaryEn":"A local Jev tool-routing gateway for coding Agents, with an MCP suggestion interface.","jevDecisionPoint":"Jev 从当前工具候选中选下一步；不同适配器可记录建议或影响实际工具选择。","jevDecisionPointEn":"Jev selects among available tools; adapters may record suggestions or influence actual tool selection.","highlightBenefit":"保留决策日志，部分客户端仅观察或自行决定是否采纳。","highlightBenefitEn":"Keeps decision logs while some clients remain observation-only or choose whether to follow suggestions.","tags":["mcp-integrations","coding-agents","llm-routing-cost"],"stars":3,"forks":1,"openIssues":0,"license":"MIT","language":"TypeScript","lastCommitAt":"2026-09-18T03:04:34Z","createdAt":"2026-09-18T02:07:34Z","metadataFetchedAt":"2026-09-21T22:01:14.962Z","avatarUrl":"https://avatars.githubusercontent.com/u/88213906?v=4","summarySource":"human-reviewed","claimStatus":"已核对固定版本 README 与调用源码。","claimStatusEn":"Reviewed the fixed-version README and call path.","plainSummaryJa":"コーディング Agent 向けのローカル Jev ツールルーター。MCP の提案インターフェースもある。","plainSummaryKo":"코딩 Agent용 로컬 Jev 도구 라우터로 MCP 제안 인터페이스도 제공한다.","jevDecisionPointJa":"候補ツールから Jev が次を選び、アダプターにより記録または実際の選択へ反映する。","jevDecisionPointKo":"Jev가 후보 도구 중 다음 것을 고르면 어댑터가 기록하거나 실제 선택에 반영한다.","highlightBenefitJa":"判断ログを保持。一部クライアントは観察のみ、または採用を自身で決める。","highlightBenefitKo":"판단 로그를 보존하며 일부 클라이언트는 관찰만 하거나 제안 수용을 스스로 결정한다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/felpsdev/jev-classifier/blob/e0c6cc7c6d2b1027be86ad8f4e03e191a88ec402/README.md","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/felpsdev/jev-classifier/blob/e0c6cc7c6d2b1027be86ad8f4e03e191a88ec402/src/jev.ts","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/felpsdev/jev-classifier/blob/e0c6cc7c6d2b1027be86ad8f4e03e191a88ec402/src/router.ts","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/felpsdev/jev-classifier/blob/e0c6cc7c6d2b1027be86ad8f4e03e191a88ec402/src/mcp.ts","note":"固定版本说明与源码；本站未执行该项目。"}],"pinned":false},{"id":"samuelsacco:jev-exploration","name":"jev-exploration","author":"SamuelSacco","url":"https://github.com/SamuelSacco/jev-exploration","category":"Evaluation & Observability","plainSummary":"记录 Jev 能力与限制的研究仓库，包含主张审查、概率校准实验和可运行示例。","plainSummaryEn":"A research repository tracking Jev claims and limitations, with calibration experiments and runnable examples.","jevDecisionPoint":"用固定问题与标注案例调用 Jev，保存回答并分析错误、校准及不同难度下的表现。","jevDecisionPointEn":"Calls Jev on defined questions and labeled cases, then analyzes errors, calibration, and difficulty effects.","highlightBenefit":"把研究结论与实验代码、数据和证据账本对应起来。","highlightBenefitEn":"Links research claims to experiment code, data, and an evidence ledger.","tags":["evaluation-benchmarks","typed-decisions"],"stars":3,"forks":0,"openIssues":4,"license":null,"language":"Python","lastCommitAt":"2026-09-18T01:45:51Z","createdAt":"2026-09-17T13:59:48Z","metadataFetchedAt":"2026-09-21T22:01:17.997Z","avatarUrl":"https://avatars.githubusercontent.com/u/76980320?v=4","summarySource":"human-reviewed","claimStatus":"作者的校准与基准结论受数据集、样本及协议限制；本站核对接入源码，未复现实验。","claimStatusEn":"The author’s calibration and benchmark conclusions are limited by datasets, samples, and protocols. This site reviewed integration code but did not reproduce the experiments.","plainSummaryJa":"Jev の主張と限界を記録し、確率校正の実験や実行例を収めた研究リポジトリ。","plainSummaryKo":"Jev의 주장과 한계를 기록하고 확률 보정 실험과 실행 예제를 담은 연구 저장소.","jevDecisionPointJa":"定義した質問とラベル付き例を使い、誤り、校正、難易度の影響を分析する。","jevDecisionPointKo":"정의된 질문과 라벨 사례를 사용해 오류, 보정, 난이도 영향을 분석한다.","highlightBenefitJa":"研究上の主張をコード、データ、根拠の台帳に結び付ける。","highlightBenefitKo":"연구 주장을 코드, 데이터, 근거 기록과 연결한다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"unconfirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/SamuelSacco/jev-exploration/blob/af91a6e09bf82c7f43d142de2bf5f1573d89b1b9/jevlab/client.py","note":"Fixed-commit source reviewed / 已核对固定版本源码"},{"url":"https://github.com/SamuelSacco/jev-exploration/blob/af91a6e09bf82c7f43d142de2bf5f1573d89b1b9/README.md","note":"Fixed-commit source reviewed / 已核对固定版本源码"}],"pinned":false},{"id":"jevbook:jevscan","name":"jevscan","author":"jevbook","url":"https://github.com/jevbook/jevscan","category":"Domain & Vertical Tools","plainSummary":"读取 EVM Token 市场特征，输出关注或回避等风险判断，提供库、CLI 和 MCP 接口。","plainSummaryEn":"An EVM Token screening tool with library, CLI and MCP interfaces for market-feature-based risk judgments.","jevDecisionPoint":"默认用本地规则计算；配置 TypeSafe key 后才把特征送给 Jev 做结构化判断。","jevDecisionPointEn":"Uses a deterministic local engine by default; a TypeSafe key enables real Jev judgments over features.","highlightBenefit":"把行情特征、判断来源和分数放在一起，便于检查风险信号。","highlightBenefitEn":"Shows features, decision provenance and scores together for inspection.","tags":["finance-quant","mcp-integrations","cli-git-gates"],"stars":2,"forks":0,"openIssues":0,"license":"MIT","language":"JavaScript","lastCommitAt":"2026-09-19T03:48:29Z","createdAt":"2026-09-18T13:58:31Z","metadataFetchedAt":"2026-09-21T22:01:20.977Z","avatarUrl":"https://avatars.githubusercontent.com/u/330919766?v=4","summarySource":"human-reviewed","claimStatus":"默认输出不是 Jev 推理；不是合约字节码审计，也不证明投资安全。","claimStatusEn":"Default output is not Jev inference. This is not a contract-bytecode audit or evidence of investment safety.","plainSummaryJa":"市場特徴に基づくリスク判断を、ライブラリー・CLI・MCP で提供する EVM Token ツール。","plainSummaryKo":"시장 특성 기반 위험 판단을 라이브러리·CLI·MCP로 제공하는 EVM Token 도구.","jevDecisionPointJa":"既定はローカル規則で、TypeSafe key を設定すると Jev 判断を使う。","jevDecisionPointKo":"기본은 로컬 규칙이며 TypeSafe key를 설정해야 실제 Jev 판단을 사용한다.","highlightBenefitJa":"特徴量・判断元・スコアを並べて確認できる。","highlightBenefitKo":"특징·판단 출처·점수를 함께 확인한다.","claimStatusJa":"固定バージョンの文書とソースを確認。実行・性能・安全性は当サイトで独立検証していません。","claimStatusKo":"고정 버전 문서와 소스를 확인했습니다. 실행·성능·안전성은 이 사이트에서 독립적으로 검증하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/jevbook/jevscan/blob/cdaf9b882cc6e221ceeaac2767fac82f584b9d88/README.md","note":"README.md"},{"url":"https://github.com/jevbook/jevscan/blob/cdaf9b882cc6e221ceeaac2767fac82f584b9d88/src/index.js","note":"src/index.js"},{"url":"https://github.com/jevbook/jevscan/blob/cdaf9b882cc6e221ceeaac2767fac82f584b9d88/src/engine.js","note":"src/engine.js"}],"pinned":false},{"id":"taruntomar122:jev-askable-arm","name":"jev-askable-arm","author":"TarunTomar122","url":"https://github.com/TarunTomar122/jev-askable-arm","category":"High-Frequency & Simulation","plainSummary":"在 ManiSkill 模拟机械臂中，让 Jev 把英文目标拆成一连串预设动作。","plainSummaryEn":"Uses Jev to chain predefined skills for English-language goals in a ManiSkill robot-arm simulation.","jevDecisionPoint":"根据仿真坐标、夹爪和物体状态，从约三十个动作原语中选择动作及目标。","jevDecisionPointEn":"Chooses a skill and target from privileged object and gripper state and about thirty primitives.","highlightBenefit":"将动作选择与 Python 的底层控制分开。","highlightBenefitEn":"Separates skill selection from Python motor control.","tags":["games-simulation","typed-decisions"],"stars":9,"forks":0,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-17T16:39:30Z","createdAt":"2026-09-17T10:01:16Z","metadataFetchedAt":"2026-09-21T22:01:23.998Z","avatarUrl":"https://avatars.githubusercontent.com/u/54112921?v=4","summarySource":"human-reviewed","claimStatus":"使用特权仿真状态而非视觉；约一秒每次选择为作者 seed 0 示例，不是实机机器人验证。 已核对固定版本 README 与集成源码。","claimStatusEn":"The robot uses privileged simulation state, not vision. About one second per pick is an author-reported seed-0 example, not a physical-robot validation. README and integration source reviewed at a fixed commit; not independently run or benchmarked by this site.","plainSummaryJa":"ManiSkill の模擬ロボットアームで、英語の目標に合わせて既定動作を Jev がつなぐ。","plainSummaryKo":"ManiSkill 로봇팔 시뮬레이션에서 영어 목표에 맞춰 Jev가 사전 정의 동작을 연결한다.","jevDecisionPointJa":"物体座標と把持状態から、約30の動作と対象を選ぶ。","jevDecisionPointKo":"객체 좌표와 집게 상태로 약 30개 동작 및 대상을 고른다.","highlightBenefitJa":"動作選択と Python の低層制御を分ける。","highlightBenefitKo":"동작 선택과 Python 저수준 제어를 분리한다.","claimStatusJa":"固定版の README と実装を確認済み。このサイトでは独立した実行・性能検証をしていません。","claimStatusKo":"고정 버전 README와 구현을 확인했습니다. 이 사이트에서 독립적인 실행·성능 검증은 하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/TarunTomar122/jev-askable-arm/blob/bef98b31a122a7033dd4bfb52867cd6ccb2d7e67/jev_robotics/common.py#L76","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/TarunTomar122/jev-askable-arm/blob/bef98b31a122a7033dd4bfb52867cd6ccb2d7e67/README.md","note":"固定版本的实现与说明 / Pinned source and documentation"}],"pinned":false},{"id":"thebous:jev-flash-review","name":"jev-flash-review","author":"TheBous","url":"https://github.com/TheBous/jev-flash-review","category":"Evaluation & Observability","plainSummary":"供编码 Agent 调用的 MCP 代码审查引擎，对提交的 diff 按规则给出结构化判断。","plainSummaryEn":"An MCP review engine that evaluates Agent-supplied diffs against explicit rules.","jevDecisionPoint":"Jev 按规则检查 diff，再从实际 hunk 中选证据位置并复核问题。","jevDecisionPointEn":"Jev checks the diff, selects evidence from actual hunks and adjudicates candidate findings.","highlightBenefit":"调用方提供 diff 与业务边界；引擎不会自行扫描仓库，评分仍需人工复核。","highlightBenefitEn":"The caller supplies diffs and task boundaries; the engine does not scan the repository itself.","tags":["coding-agents","mcp-integrations","evaluation-benchmarks"],"stars":1,"forks":0,"openIssues":0,"license":null,"language":"TypeScript","lastCommitAt":"2026-09-19T12:25:52Z","createdAt":"2026-09-17T21:11:52Z","metadataFetchedAt":"2026-09-21T22:01:27.215Z","avatarUrl":"https://avatars.githubusercontent.com/u/17549100?v=4","summarySource":"human-reviewed","claimStatus":"已核对固定版本 README 与调用源码。","claimStatusEn":"Reviewed the fixed-version README and call path.","plainSummaryJa":"Agent が渡す diff を明示ルールで評価する MCP コードレビューエンジン。","plainSummaryKo":"Agent가 전달한 diff를 명시적 규칙으로 평가하는 MCP 코드 리뷰 엔진.","jevDecisionPointJa":"Jev が diff を点検し、実在する hunk から証拠位置を選んで指摘を再確認する。","jevDecisionPointKo":"Jev가 diff를 검사하고 실제 hunk에서 증거 위치를 선택해 지적 사항을 재확인한다.","highlightBenefitJa":"diff と業務境界は呼び出し側が渡し、エンジン自身はリポジトリを走査しない。","highlightBenefitKo":"diff와 업무 범위는 호출자가 제공하며 엔진이 저장소를 직접 스캔하지 않는다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"unconfirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/TheBous/jev-flash-review/blob/658a7bd52e30292daa21ca3941179429634371eb/README.md","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/TheBous/jev-flash-review/blob/658a7bd52e30292daa21ca3941179429634371eb/src/judge.ts","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/TheBous/jev-flash-review/blob/658a7bd52e30292daa21ca3941179429634371eb/src/mcp/reviewServer.ts","note":"固定版本说明与源码；本站未执行该项目。"}],"pinned":false},{"id":"kt3k:jevchat","name":"jevchat","author":"kt3k","url":"https://github.com/kt3k/jevchat","category":"Decision Tools","plainSummary":"一个聊天式 Jev 演示：回答从预先定义或自定义的选项里选择，而不是生成长文。","plainSummaryEn":"A chat-style Jev demo whose answers are selected from predefined or custom options rather than generated prose.","jevDecisionPoint":"把回答风格映射成 Choice 选项，也从问题片段中选择聊天标题。","jevDecisionPointEn":"Maps answer styles to Choice options and selects chat titles from fragments of the question.","highlightBenefit":"在聊天界面查看选项及其概率，并尝试自己的答案集合。","highlightBenefitEn":"Shows answer options and probabilities in a chat UI with customizable answer sets.","tags":["voice-conversation","typed-decisions"],"stars":1,"forks":0,"openIssues":0,"license":null,"language":"JavaScript","lastCommitAt":"2026-09-18T10:02:57Z","createdAt":"2026-09-18T01:44:16Z","metadataFetchedAt":"2026-09-21T22:01:30.075Z","avatarUrl":"https://avatars.githubusercontent.com/u/613956?v=4","summarySource":"human-reviewed","claimStatus":"聊天外观不代表自由文本生成能力；这是演示，模型选择不保证事实正确。","claimStatusEn":"The chat appearance does not imply free-text generation. This is a demo; selected answers are not guaranteed factual, and this site has not run it.","plainSummaryJa":"定義済みまたは独自の選択肢から回答する、チャット形式の Jev デモ。","plainSummaryKo":"미리 정의하거나 직접 만든 선택지에서 답을 고르는 채팅형 Jev 데모.","jevDecisionPointJa":"回答スタイルを Choice に変換し、質問の断片からチャットタイトルも選ぶ。","jevDecisionPointKo":"답변 스타일을 Choice로 변환하고 질문 조각에서 대화 제목도 선택한다.","highlightBenefitJa":"チャット画面で選択肢と確率を確認し、独自の回答集合を試せる。","highlightBenefitKo":"채팅 화면에서 선택지와 확률을 보고 직접 만든 답변 집합을 시험할 수 있다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"unconfirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/kt3k/jevchat/blob/18ecea877ebcdaa66b0d2eba9a4db87f0b9715ab/main.ts","note":"Fixed-commit source reviewed / 已核对固定版本源码"},{"url":"https://github.com/kt3k/jevchat/blob/18ecea877ebcdaa66b0d2eba9a4db87f0b9715ab/README.md","note":"Fixed-commit source reviewed / 已核对固定版本源码"}],"pinned":false},{"id":"raihankhan-rk:jevarena","name":"jevarena","author":"raihankhan-rk","url":"https://github.com/raihankhan-rk/jevarena","category":"High-Frequency & Simulation","plainSummary":"两个 Jev Agent 在并排的浏览器 Snake 游戏里对战，观众可以查看每步方向选择。","plainSummaryEn":"Two Jev Agents play Snake in side-by-side browser panes with visible per-step choices.","jevDecisionPoint":"根据结构化棋盘状态，从当前允许的方向按钮中选择点击目标。","jevDecisionPointEn":"Selects a direction-button target from allowed controls using structured board state.","highlightBenefit":"把候选按钮、操作概率和游戏进展一起显示。","highlightBenefitEn":"Displays candidate controls, action probabilities and game progress together.","tags":["games-simulation","typed-decisions"],"stars":2,"forks":0,"openIssues":0,"license":"MIT","language":"TypeScript","lastCommitAt":"2026-09-18T10:03:15Z","createdAt":"2026-09-18T06:38:12Z","metadataFetchedAt":"2026-09-21T22:01:32.979Z","avatarUrl":"https://avatars.githubusercontent.com/u/34377027?v=4","summarySource":"human-reviewed","claimStatus":"Jev 接收结构化状态而非截图；演示及离线路径不能当作实时模型胜率证据。","claimStatusEn":"Jev receives structured state rather than screenshots; demo or offline paths do not establish live-model win rates. Not run here.","plainSummaryJa":"二つのブラウザー画面で Jev Agent が Snake をプレイし、一手ごとの選択を表示する。","plainSummaryKo":"나란한 두 브라우저 화면에서 Jev Agent가 Snake를 하며 단계별 선택을 표시한다.","jevDecisionPointJa":"構造化した盤面状態から、許可された方向ボタンを選ぶ。","jevDecisionPointKo":"구조화된 보드 상태로 허용된 방향 버튼을 선택한다.","highlightBenefitJa":"候補・操作確率・ゲーム進行を並べて確認できる。","highlightBenefitKo":"후보·동작 확률·게임 진행을 함께 확인한다.","claimStatusJa":"固定バージョンの文書とソースを確認。実行・性能・安全性は当サイトで独立検証していません。","claimStatusKo":"고정 버전 문서와 소스를 확인했습니다. 실행·성능·안전성은 이 사이트에서 독립적으로 검증하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/raihankhan-rk/jevarena/blob/c22adde05edfc3ee16911d38b12b337c5f9727a1/README.md","note":"README.md"},{"url":"https://github.com/raihankhan-rk/jevarena/blob/c22adde05edfc3ee16911d38b12b337c5f9727a1/lib/server/jev.ts","note":"lib/server/jev.ts"}],"pinned":false},{"id":"supercorp-ai:supercov","name":"supercov","author":"supercorp-ai","url":"https://github.com/supercorp-ai/supercov","category":"Evaluation & Observability","plainSummary":"编程 Agent 的代码质量与覆盖率 CLI：Jev 检查源码属性，本地覆盖率工具帮助选择补测目标。","plainSummaryEn":"A quality and coverage CLI for coding agents: Jev assesses source properties while local coverage highlights testing targets.","jevDecisionPoint":"向 Jev 询问每个文件的质量属性，由程序汇总分数与优先顺序。","jevDecisionPointEn":"Asks Jev about file-quality properties; code composes scores and ordering.","highlightBenefit":"把分数拆成可对照源码的命名属性，并缓存按内容得到的结果。","highlightBenefitEn":"Breaks scores into named properties and caches results by content.","tags":["coding-agents","cli-git-gates","evaluation-benchmarks"],"stars":90,"forks":4,"openIssues":1,"license":"MIT","lastCommitAt":"2026-09-21T17:58:21Z","createdAt":"2026-08-23T16:12:35Z","metadataFetchedAt":"2026-09-21T22:01:35.962Z","avatarUrl":"https://avatars.githubusercontent.com/u/152935339?v=4","summarySource":"human-reviewed","claimStatus":"已核对固定版本 README 与集成源码。","claimStatusEn":"README and integration source reviewed at a fixed commit; not independently run or benchmarked by this site.","plainSummaryJa":"Jev によるコード属性確認と、ローカルのカバレッジを組み合わせた Agent 向け CLI。","plainSummaryKo":"Jev 코드 속성 검사와 로컬 커버리지를 결합한 코딩 Agent용 CLI다.","jevDecisionPointJa":"ファイルの品質属性を Jev に問い、コードがスコアと順序をまとめる。","jevDecisionPointKo":"파일 품질 속성을 Jev에 질문하고 코드가 점수와 순서를 만든다.","highlightBenefitJa":"スコアを名前付き属性に分け、内容ごとに結果を保存する。","highlightBenefitKo":"점수를 이름 붙인 속성으로 나누고 내용별 결과를 캐시한다.","claimStatusJa":"固定版の README と実装を確認済み。このサイトでは独立した実行・性能検証をしていません。","claimStatusKo":"고정 버전 README와 구현을 확인했습니다. 이 사이트에서 독립적인 실행·성능 검증은 하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/supercorp-ai/supercov/blob/55f5ce93a239829c224b89e6749991310be91ea4/crates/supercov-cli/src/quality.rs#L30","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/supercorp-ai/supercov/blob/55f5ce93a239829c224b89e6749991310be91ea4/README.md","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/supercorp-ai/supercov/blob/55f5ce93a239829c224b89e6749991310be91ea4/contracts/v1/README.md","note":"固定版本的实现与说明 / Pinned source and documentation"}],"pinned":false},{"id":"doeixd:jev-pref","name":"jev-pref","author":"doeixd","url":"https://github.com/doeixd/jev-pref","category":"Evaluation & Observability","plainSummary":"把 AGENTS.md 中的项目偏好整理为规则，再用 Jev 检查 hunk、暂存文件或 PR。","plainSummaryEn":"Turns AGENTS.md preferences into rules checked by Jev against hunks, staged files or PRs.","jevDecisionPoint":"Jev 判断代码变更是否触及配置规则，程序将答案映射为检查结果。","jevDecisionPointEn":"Jev classifies change evidence against configured rules; code maps answers to review outcomes.","highlightBenefit":"可把语义规则反馈给编码 Agent，不能替代类型检查、测试或安全审计。","highlightBenefitEn":"Feeds semantic-rule findings to coding Agents without replacing type checks, tests or security audits.","tags":["coding-agents","cli-git-gates","evaluation-benchmarks"],"stars":4,"forks":0,"openIssues":1,"license":"MIT","language":"JavaScript","lastCommitAt":"2026-09-18T17:49:07Z","createdAt":"2026-09-17T18:01:44Z","metadataFetchedAt":"2026-09-21T22:01:38.950Z","avatarUrl":"https://avatars.githubusercontent.com/u/13461122?v=4","summarySource":"human-reviewed","claimStatus":"已核对固定版本 README 与调用源码。","claimStatusEn":"Reviewed the fixed-version README and call path.","plainSummaryJa":"AGENTS.md のプロジェクト方針をルールに整理し、hunk・ステージ済みファイル・PR を Jev で点検する。","plainSummaryKo":"AGENTS.md의 프로젝트 선호를 규칙으로 정리해 hunk·스테이징 파일·PR을 Jev로 검사한다.","jevDecisionPointJa":"変更証拠を設定ルールに照らして分類し、コードが結果へ変換する。","jevDecisionPointKo":"변경 증거를 설정 규칙에 따라 분류하면 코드가 검토 결과로 변환한다.","highlightBenefitJa":"意味ルールの指摘を Agent に返すが、型検査・テスト・安全監査の代替ではない。","highlightBenefitKo":"의미 규칙 지적을 Agent에 전달하지만 타입 검사·테스트·보안 감사를 대체하지 않는다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/doeixd/jev-pref/blob/9d77ea6069123a0b3eedf465ec39781842722b0a/README.md","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/doeixd/jev-pref/blob/9d77ea6069123a0b3eedf465ec39781842722b0a/packages/jev-pref/src/jev.js","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/doeixd/jev-pref/blob/9d77ea6069123a0b3eedf465ec39781842722b0a/packages/jev-pref/src/commands/review.js","note":"固定版本说明与源码；本站未执行该项目。"}],"pinned":false,"topics":["agent-skills","claude-code","code-review","jev","skills-sh","typesafe"]},{"id":"kylemclaren:jevql","name":"jevql","author":"kylemclaren","url":"https://github.com/kylemclaren/jevql","category":"Data & Search","plainSummary":"在普通 PostgreSQL 查询外加一层语义处理，用 Jev 筛选、分组和排序，无需安装数据库扩展。","plainSummaryEn":"Adds Jev semantic filtering, grouping, and ranking around ordinary PostgreSQL without installing a database extension.","jevDecisionPoint":"CLI 或服务层解析 jev_* 调用，把行文本发给 Jev，再用结果完成查询。","jevDecisionPointEn":"The CLI or service layer evaluates jev_* calls over row text before completing the query.","highlightBenefit":"通过 CLI、HTTP、MCP 与 SDK 复用同一套语义 SQL 接口。","highlightBenefitEn":"Shares semantic SQL interfaces across a CLI, HTTP, MCP, and SDKs.","tags":["database-vector","classification-ranking"],"stars":11,"forks":0,"openIssues":2,"license":"MIT","language":"Go","lastCommitAt":"2026-09-19T11:12:01Z","createdAt":"2026-09-18T11:57:33Z","metadataFetchedAt":"2026-09-21T22:01:41.985Z","avatarUrl":"https://avatars.githubusercontent.com/u/3727384?v=4","summarySource":"human-reviewed","claimStatus":"Jev 调用发生在客户端或服务层，不是 Postgres 内核的原生语义执行；行内容会发给外部 API。","claimStatusEn":"Jev calls run in the client or service layer, not inside the Postgres engine. Row content is sent to an external API.","plainSummaryJa":"データベース拡張を入れずに、通常の PostgreSQL へ Jev による絞り込み、分類、順位付けを加える。","plainSummaryKo":"데이터베이스 확장 설치 없이 일반 PostgreSQL에 Jev 기반 필터링, 분류, 순위를 추가한다.","jevDecisionPointJa":"CLI やサービス層が jev_* 呼び出しを解析し、行テキストの判断結果で問い合わせを処理する。","jevDecisionPointKo":"CLI나 서비스 계층이 jev_* 호출을 해석하고 행 텍스트 판단 결과로 쿼리를 처리한다.","highlightBenefitJa":"CLI、HTTP、MCP、SDK で同じ意味検索 SQL を使える。","highlightBenefitKo":"CLI, HTTP, MCP, SDK에서 같은 의미 기반 SQL을 사용할 수 있다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"confirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/kylemclaren/jevql/blob/b5e546bafd18a2c887142bbbdfbce7ed9e06960e/internal/typesafe/client.go","note":"Fixed-commit source reviewed / 已核对固定版本源码"},{"url":"https://github.com/kylemclaren/jevql/blob/b5e546bafd18a2c887142bbbdfbce7ed9e06960e/README.md","note":"Fixed-commit source reviewed / 已核对固定版本源码"}],"pinned":false,"topics":["jev","postgres"]},{"id":"vercel-labs:json-render","name":"json-render","author":"vercel-labs","url":"https://github.com/vercel-labs/json-render","category":"Creative Tools","plainSummary":"json-render 网站里的 Jev UI 组合实验：从预定义组件与属性候选中选择，再由代码组装界面。","plainSummaryEn":"A Jev UI-composition experiment inside the json-render website that selects from predefined component and property options.","jevDecisionPoint":"通过 Vercel AI Gateway 评估组件配置，composeSpec 把答案变成界面规格。","jevDecisionPointEn":"Evaluates component configurations through Vercel AI Gateway, then composeSpec assembles the UI specification.","highlightBenefit":"提供一条与逐 Token 生成 JSON 不同的、可检查的界面组合路径。","highlightBenefitEn":"Provides an inspectable composition path alongside token-by-token JSON generation.","tags":["creative-multimedia","typed-decisions"],"stars":17921,"forks":938,"openIssues":109,"license":"Apache-2.0","language":"TypeScript","lastCommitAt":"2026-09-18T18:44:50Z","createdAt":"2026-01-14T17:22:39Z","metadataFetchedAt":"2026-09-21T22:01:44.992Z","avatarUrl":"https://avatars.githubusercontent.com/u/108547162?v=4","summarySource":"human-reviewed","claimStatus":"仅指仓库网站中的 Jev 实验路径；未独立验证 3.21 秒对 880 毫秒等性能说法，也不表示整个框架依赖 Jev。","claimStatusEn":"This covers the website’s Jev experiment, not the whole framework. Claims such as 3.21 s versus 880 ms were not independently verified.","plainSummaryJa":"json-render のサイト内で、定義済みコンポーネントと属性を Jev が選ぶ UI 組み立て実験。","plainSummaryKo":"json-render 사이트에서 미리 정의한 컴포넌트와 속성을 Jev가 고르는 UI 구성 실험.","jevDecisionPointJa":"Vercel AI Gateway で構成を評価し、composeSpec が UI 仕様にまとめる。","jevDecisionPointKo":"Vercel AI Gateway로 구성을 평가하고 composeSpec이 UI 명세를 조립한다.","highlightBenefitJa":"Token ごとの JSON 生成とは別の、確認可能な構成経路を提供する。","highlightBenefitKo":"Token별 JSON 생성과 별도로 확인 가능한 구성 경로를 제공한다.","claimStatusJa":"固定バージョンの文書とソースを確認。実行・性能・安全性は当サイトで独立検証していません。","claimStatusKo":"고정 버전 문서와 소스를 확인했습니다. 실행·성능·안전성은 이 사이트에서 독립적으로 검증하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/vercel-labs/json-render/blob/3ad381881194e7011ad3ccd6d668033495a06c29/README.md","note":"README.md"},{"url":"https://github.com/vercel-labs/json-render/blob/3ad381881194e7011ad3ccd6d668033495a06c29/apps/web/lib/jev/compose.ts","note":"apps/web/lib/jev/compose.ts"}],"pinned":false},{"id":"trycua:cua","name":"cua","author":"trycua","url":"https://github.com/trycua/cua","category":"Browser & OS Action","plainSummary":"Cua 仓库的 jev-use 预览示例：Driver 观察与执行，Jev 从有界候选中选择浏览器动作。","plainSummaryEn":"Cua’s preview jev-use example pairs Driver observation and execution with bounded Jev browser-action choices.","jevDecisionPoint":"读取 DOM 或可用的视觉区域描述，仅返回已提供的候选动作 ID。","jevDecisionPointEn":"Reads DOM or supported visual-region descriptions and returns a supplied candidate action ID.","highlightBenefit":"提供 Python、TypeScript 循环，以及离线与在线分开的验证路径。","highlightBenefitEn":"Provides Python and TypeScript loops with separate offline and live checks.","tags":["browser-automation","typed-decisions"],"stars":25658,"forks":1766,"openIssues":1036,"license":"MIT","language":"Rust","lastCommitAt":"2026-09-19T17:37:04Z","createdAt":"2025-01-31T15:02:49Z","metadataFetchedAt":"2026-09-21T22:01:47.979Z","avatarUrl":"https://avatars.githubusercontent.com/u/191107687?v=4","summarySource":"human-reviewed","claimStatus":"Jev 集成位于预览示例，不能代表整个 Cua 框架或 CUA-S1；离线 fixture 通过不代表线上模型效果。 已核对固定版本 README 与集成源码。","claimStatusEn":"The Jev integration is a public-preview example, not the entire Cua framework or CUA-S1. Offline fixture checks do not establish live model performance. README and integration source reviewed at a fixed commit; not independently run or benchmarked by this site.","plainSummaryJa":"Cua の試験的な jev-use 例が Driver の観察・実行と Jev の候補選択を組み合わせる。","plainSummaryKo":"Cua의 미리보기 jev-use 예제가 Driver 관찰·실행과 Jev의 제한된 동작 선택을 연결한다.","jevDecisionPointJa":"DOM または対応する視覚領域の説明を読み、提示済みの操作 ID を返す。","jevDecisionPointKo":"DOM 또는 지원되는 시각 영역 설명을 읽고 제시된 동작 ID를 반환한다.","highlightBenefitJa":"Python と TypeScript のループと、オフライン・実 API の別々の検証経路を備える。","highlightBenefitKo":"Python·TypeScript 루프와 오프라인·실제 API 검증 경로를 제공한다.","claimStatusJa":"固定版の README と実装を確認済み。このサイトでは独立した実行・性能検証をしていません。","claimStatusKo":"고정 버전 README와 구현을 확인했습니다. 이 사이트에서 독립적인 실행·성능 검증은 하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/trycua/cua/blob/83f142c4290a0f7d9ed545ae8532858c6e4f8145/libs/cua-driver/examples/jev-use/python/jev_adapter.py#L11","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/trycua/cua/blob/83f142c4290a0f7d9ed545ae8532858c6e4f8145/libs/cua-driver/examples/jev-use/README.md","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/trycua/cua/blob/83f142c4290a0f7d9ed545ae8532858c6e4f8145/README.md","note":"固定版本的实现与说明 / Pinned source and documentation"}],"pinned":false,"topics":["computer-use","automation","agent","system-one","jev"]},{"id":"abeatrix:cline-plugin-jev-browser","name":"cline-plugin-jev-browser","author":"abeatrix","url":"https://github.com/abeatrix/cline-plugin-jev-browser","category":"Browser & OS Action","plainSummary":"Cline 的独立 Playwright 浏览器插件，通过 Vercel AI Gateway 用 Jev 选择网页操作。","plainSummaryEn":"A Cline plugin using an isolated Playwright browser and Jev decisions through Vercel AI Gateway.","jevDecisionPoint":"Jev 读取 DOM 目标表选择动作；需输入文字时由另外的文本模型生成。","jevDecisionPointEn":"Jev selects operations from a DOM target table; a separate text model supplies typed text when needed.","highlightBenefit":"保存操作前后截图并在敏感动作前交还控制；完成状态仍需回读验证。","highlightBenefitEn":"Captures before-and-after screenshots and yields on sensitive actions; completion still needs readback verification.","tags":["browser-automation","coding-agents"],"stars":0,"forks":0,"openIssues":0,"license":null,"language":"TypeScript","lastCommitAt":"2026-09-18T20:00:28Z","createdAt":"2026-09-18T06:43:54Z","metadataFetchedAt":"2026-09-21T22:01:50.966Z","avatarUrl":"https://avatars.githubusercontent.com/u/68532117?v=4","summarySource":"human-reviewed","claimStatus":"已核对固定版本 README 与调用源码。","claimStatusEn":"Reviewed the fixed-version README and call path.","plainSummaryJa":"独立した Playwright ブラウザーと Vercel AI Gateway 経由の Jev 判断を使う Cline プラグイン。","plainSummaryKo":"격리된 Playwright 브라우저와 Vercel AI Gateway의 Jev 판단을 사용하는 Cline 플러그인.","jevDecisionPointJa":"DOM 対象表から Jev が操作を選び、必要な入力文は別のテキストモデルが作る。","jevDecisionPointKo":"Jev가 DOM 대상 표에서 동작을 선택하며 입력 문장은 별도 텍스트 모델이 만든다.","highlightBenefitJa":"前後のスクリーンショットを保存し、重要操作は制御を返す。完了は結果確認が必要。","highlightBenefitKo":"전후 스크린샷을 저장하고 민감한 동작은 제어를 반환하며 완료 결과는 재확인이 필요하다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"unconfirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/abeatrix/cline-plugin-jev-browser/blob/c886dcb35cb7d64d682a56211f9b04a5345c6433/README.md","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/abeatrix/cline-plugin-jev-browser/blob/c886dcb35cb7d64d682a56211f9b04a5345c6433/src/jev-model.ts","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/abeatrix/cline-plugin-jev-browser/blob/c886dcb35cb7d64d682a56211f9b04a5345c6433/src/jev-run.ts","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/abeatrix/cline-plugin-jev-browser/blob/c886dcb35cb7d64d682a56211f9b04a5345c6433/index.ts","note":"固定版本说明与源码；本站未执行该项目。"}],"pinned":false},{"id":"virattt:ai-hedge-fund","name":"ai-hedge-fund","author":"virattt","url":"https://github.com/virattt/ai-hedge-fund","category":"Domain & Vertical Tools","plainSummary":"一个教育用途的 AI 对冲基金原型，其中可选 Jev 适配器把结构化判断接到基金决策流程。","plainSummaryEn":"An educational AI hedge-fund prototype with an optional Jev adapter for structured judgments in fund decision workflows.","jevDecisionPoint":"把策略问题转为 System One 请求，再把原生答案转换成项目统一的结果格式。","jevDecisionPointEn":"Converts strategy questions to System One requests and normalizes native answers to the project’s result format.","highlightBenefit":"在同一研究流程里选择使用 Jev 或其他模型后端。","highlightBenefitEn":"Allows Jev or other model backends to be used in the same research workflow.","tags":["finance-quant","typed-decisions"],"stars":63654,"forks":11158,"openIssues":166,"license":"MIT","language":"Python","lastCommitAt":"2026-09-18T14:54:20Z","createdAt":"2024-11-29T16:30:01Z","metadataFetchedAt":"2026-09-21T22:01:53.953Z","avatarUrl":"https://avatars.githubusercontent.com/u/901795?v=4","summarySource":"human-reviewed","claimStatus":"作者明确当前是教育研究原型、并不实际交易；不能称“官方 Jev 适配器”“零幻觉”或宣传盈利能力。本站未回测。","claimStatusEn":"The author describes an educational research prototype that currently does not execute trades. It is not an official Jev adapter, a zero-hallucination system, or evidence of profitability. This site has not backtested it.","plainSummaryJa":"基金判断の流れに任意の Jev アダプターを持つ、教育用途の AI ヘッジファンド試作。","plainSummaryKo":"펀드 판단 과정에 선택적 Jev 어댑터를 둔 교육용 AI 헤지펀드 프로토타입.","jevDecisionPointJa":"戦略上の質問を System One に送り、回答をプロジェクト共通の形式に変換する。","jevDecisionPointKo":"전략 질문을 System One 요청으로 보내고 답을 프로젝트 공통 형식으로 변환한다.","highlightBenefitJa":"同じ研究フローで Jev やほかのモデルを選べる。","highlightBenefitKo":"동일한 연구 흐름에서 Jev나 다른 모델을 선택할 수 있다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"confirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/virattt/ai-hedge-fund/blob/154a8b2f46dca0f40764d814e4e747b0ad71f4c4/hedge_fund/llm/client.py","note":"Fixed-commit source reviewed / 已核对固定版本源码"},{"url":"https://github.com/virattt/ai-hedge-fund/blob/154a8b2f46dca0f40764d814e4e747b0ad71f4c4/README.md","note":"Fixed-commit source reviewed / 已核对固定版本源码"}],"pinned":false},{"id":"davila7:claude-code-templates","name":"jev-model-router","author":"davila7","url":"https://github.com/davila7/claude-code-templates","category":"Routing & Cost Optimization","plainSummary":"claude-code-templates 社区仓库中的 Jev 路由模组，为 Claude Code 子 Agent 建议模型与思考档位。","plainSummaryEn":"A community Claude Code Templates mod that uses Jev to suggest subagent models and reasoning levels.","jevDecisionPoint":"评估任务级别、推理需求和生产风险，由本地策略映射成调用配置。","jevDecisionPointEn":"Evaluates task tier, reasoning needs and production risk; local policy maps results to invocation settings.","highlightBenefit":"把模型选择规则单独配置，便于核对每次路由原因。","highlightBenefitEn":"Makes routing rules and per-request choices explicit and configurable.","tags":["llm-routing-cost","coding-agents"],"stars":30889,"forks":3521,"openIssues":261,"license":"MIT","language":"Python","lastCommitAt":"2026-09-21T21:35:58Z","createdAt":"2025-07-04T01:26:26Z","metadataFetchedAt":"2026-09-21T22:01:56.949Z","avatarUrl":"https://avatars.githubusercontent.com/u/6216945?v=4","summarySource":"human-reviewed","claimStatus":"davila7 的社区项目，不是 Anthropic 官方仓库或官方模组；本站未验证综合降本或安全效果。","claimStatusEn":"A davila7 community project, not an official Anthropic repository or mod. Aggregate savings and safety effects were not verified here.","plainSummaryJa":"Jev で Claude Code の子 Agent のモデル・思考レベルを提案するコミュニティ製 mod。","plainSummaryKo":"Jev로 Claude Code 하위 Agent의 모델·추론 수준을 제안하는 커뮤니티 mod.","jevDecisionPointJa":"タスク段階・推論要件・本番リスクを評価し、ローカル方針で設定へ変換する。","jevDecisionPointKo":"작업 수준·추론 필요·운영 위험을 평가하고 로컬 정책이 호출 설정으로 변환한다.","highlightBenefitJa":"振り分け規則と各選択の理由を確認・設定できる。","highlightBenefitKo":"라우팅 규칙과 요청별 선택을 확인하고 설정할 수 있다.","claimStatusJa":"固定バージョンの文書とソースを確認。実行・性能・安全性は当サイトで独立検証していません。","claimStatusKo":"고정 버전 문서와 소스를 확인했습니다. 실행·성능·안전성은 이 사이트에서 독립적으로 검증하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/davila7/claude-code-templates/blob/61bfcd1586bf1076f6d3cfa0436317c912811e6c/README.md","note":"README.md"},{"url":"https://github.com/davila7/claude-code-templates/blob/61bfcd1586bf1076f6d3cfa0436317c912811e6c/scripts/jev-spike.mjs","note":"scripts/jev-spike.mjs"},{"url":"https://github.com/davila7/claude-code-templates/blob/61bfcd1586bf1076f6d3cfa0436317c912811e6c/cli-tool/components/mods/productivity/jev-model-router/hooks/jev-model-router.ts","note":"cli-tool/components/mods/productivity/jev-model-router/hooks/jev-model-router.ts"},{"url":"https://github.com/davila7/claude-code-templates/blob/61bfcd1586bf1076f6d3cfa0436317c912811e6c/cli-tool/components/mods/productivity/jev-model-router/README.md","note":"cli-tool/components/mods/productivity/jev-model-router/README.md"}],"pinned":false,"topics":["anthropic","anthropic-claude","claude","claude-code"]},{"id":"0xplaygrounds:rig","name":"rig-typesafeai","author":"0xPlaygrounds","url":"https://github.com/0xPlaygrounds/rig","category":"SDK & Decision Frameworks","plainSummary":"Rig 仓库中的实验性 TypeSafe crate，用 Rust 类型组织 Jev 的问题与答案。","plainSummaryEn":"An experimental TypeSafe crate in Rig for expressing Jev questions and answers with Rust types.","jevDecisionPoint":"把应用状态和问题发送给 Jev，解析 Choice、Score 或 Noul 答案。","jevDecisionPointEn":"Sends application state and questions to Jev and parses Choice, Score or Noul answers.","highlightBenefit":"复用问题与答案的字段布局，并验证答案是否符合问题。","highlightBenefitEn":"Reuses field layouts between questions and answers and validates answers against questions.","tags":["multilanguage-sdk","typed-decisions"],"stars":8692,"forks":974,"openIssues":103,"license":"MIT","language":"Rust","lastCommitAt":"2026-09-21T21:45:12Z","createdAt":"2024-06-05T13:42:28Z","metadataFetchedAt":"2026-09-21T22:02:00.003Z","avatarUrl":"https://avatars.githubusercontent.com/u/93353392?v=4","summarySource":"human-reviewed","claimStatus":"已核对固定版本 README 与集成源码。","claimStatusEn":"README and integration source reviewed at a fixed commit; not independently run or benchmarked by this site.","plainSummaryJa":"Rust 型で Jev の質問と回答を構成する、Rig 内の実験的な TypeSafe crate。","plainSummaryKo":"Rust 타입으로 Jev 질문과 답변을 구성하는 Rig의 실험적 TypeSafe crate다.","jevDecisionPointJa":"状態と質問を Jev に送り、Choice、Score、Noul の回答を解析する。","jevDecisionPointKo":"상태와 질문을 Jev에 보내 Choice, Score, Noul 답변을 처리한다.","highlightBenefitJa":"質問と回答でフィールド構成を再利用し、対応する回答を検証する。","highlightBenefitKo":"질문과 답변의 필드 구조를 재사용하고 답변이 질문에 맞는지 검사한다.","claimStatusJa":"固定版の README と実装を確認済み。このサイトでは独立した実行・性能検証をしていません。","claimStatusKo":"고정 버전 README와 구현을 확인했습니다. 이 사이트에서 독립적인 실행·성능 검증은 하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/0xPlaygrounds/rig/blob/2d16c1b25f6749b3a2cd841beddf767106495069/crates/rig-typesafeai/src/wire.rs#L40","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/0xPlaygrounds/rig/blob/2d16c1b25f6749b3a2cd841beddf767106495069/crates/rig-typesafeai/README.md","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/0xPlaygrounds/rig/blob/2d16c1b25f6749b3a2cd841beddf767106495069/examples/typesafeai_chat/README.md","note":"固定版本的实现与说明 / Pinned source and documentation"}],"pinned":false,"topics":["rust","rig","llm","typesafe","framework"]},{"id":"jarrodwatts:jev-trader","name":"jev-trader","author":"jarrodwatts","url":"https://github.com/jarrodwatts/jev-trader","category":"Domain & Vertical Tools","plainSummary":"在 Monad 的 Kuru MON-USDC 订单簿上尝试做市，Jev 可按新区块选择买卖方向。","plainSummaryEn":"A market-making experiment on Kuru’s MON-USDC book on Monad, with optional per-block Jev buy/sell decisions.","jevDecisionPoint":"开启 Jev 模式后读取订单簿并选择方向，代码模拟成交或按配置提交 post-only 限价单。","jevDecisionPointEn":"In Jev mode, order-book judgments feed code that simulates fills or submits configured post-only limit orders.","highlightBenefit":"默认 mock 模型；无私钥时为 dry run，模拟结果不证明可获利。","highlightBenefitEn":"Defaults to a mock model and dry-runs without a private key; simulated results do not establish profitability.","tags":["finance-quant","typed-decisions"],"stars":1827,"forks":350,"openIssues":4,"license":"MIT","language":"TypeScript","lastCommitAt":"2026-09-17T02:48:21Z","createdAt":"2026-09-16T21:36:20Z","metadataFetchedAt":"2026-09-21T22:02:02.966Z","avatarUrl":"https://avatars.githubusercontent.com/u/35651410?v=4","summarySource":"human-reviewed","claimStatus":"默认 mock；无钱包私钥时模拟成交。未验证实盘表现、延迟或收益。","claimStatusEn":"Defaults to a mock model and simulates fills without a wallet key. Live trading, latency and profitability were not verified.","plainSummaryJa":"Monad の Kuru MON-USDC 板で、Jev のブロック単位判断を選択できるマーケットメイク実験。","plainSummaryKo":"Monad Kuru MON-USDC 호가창에서 블록별 Jev 매수·매도 판단을 선택할 수 있는 시장조성 실험.","jevDecisionPointJa":"Jev モードでは板から売買方向を選び、コードが模擬約定または設定済み指値注文を扱う。","jevDecisionPointKo":"Jev 모드의 방향 판단으로 코드가 모의 체결 또는 설정된 post-only 지정가 주문을 처리한다.","highlightBenefitJa":"既定は mock モデル。秘密鍵なしでは dry run で、模擬結果は収益性を証明しない。","highlightBenefitKo":"기본은 mock 모델이며 개인 키가 없으면 dry run으로 실행돼 수익성을 증명하지 않는다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/jarrodwatts/jev-trader/blob/b587759e459ea049590102e54a0b07800864cdc3/README.md","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/jarrodwatts/jev-trader/blob/b587759e459ea049590102e54a0b07800864cdc3/src/model.ts","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/jarrodwatts/jev-trader/blob/b587759e459ea049590102e54a0b07800864cdc3/src/trader.ts","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/jarrodwatts/jev-trader/blob/b587759e459ea049590102e54a0b07800864cdc3/src/config.ts","note":"固定版本说明与源码；本站未执行该项目。"}],"pinned":false},{"id":"awlevin:typesafe-computer-use","name":"typesafe-computer-use","author":"awlevin","url":"https://github.com/awlevin/typesafe-computer-use","category":"Browser & OS Action","plainSummary":"用 OCR 和界面状态构造候选动作，让 Jev 决定如何操作 macOS，需要写文字时再调用文本模型。","plainSummaryEn":"Builds candidate actions from OCR and UI state for Jev to control macOS, calling a text model only when writing is needed.","jevDecisionPoint":"从确定性提取的控件与动作中选择下一步，执行器操作桌面。","jevDecisionPointEn":"Selects the next step from deterministically extracted controls and actions before desktop execution.","highlightBenefit":"把屏幕读取、动作选择与文本生成分开。","highlightBenefitEn":"Separates screen reading, action selection, and text generation.","tags":["desktop-os","typed-decisions"],"stars":745,"forks":59,"openIssues":11,"license":"MIT","language":"Python","lastCommitAt":"2026-09-18T23:33:48Z","createdAt":"2026-09-16T17:48:45Z","metadataFetchedAt":"2026-09-21T22:02:05.948Z","avatarUrl":"https://avatars.githubusercontent.com/u/8488209?v=4","summarySource":"human-reviewed","claimStatus":"README 的成本与速度对照来自作者限定截图与目标的测量，且部分推理由确定性代码补足；不能推广为通用降本倍数。","claimStatusEn":"README cost and speed comparisons use the author’s limited screenshot and goal setup, with some reasoning supplied by deterministic code. They are not general savings multipliers and were not independently retested.","plainSummaryJa":"OCR と画面状態から候補を作り、Jev が macOS の操作を選ぶ。文章入力時は別モデルを使う。","plainSummaryKo":"OCR과 화면 상태에서 후보를 만들고 Jev가 macOS 동작을 선택하며 문장 입력은 별도 모델을 사용한다.","jevDecisionPointJa":"抽出した要素と操作候補から次の一手を選び、実行器がデスクトップを操作する。","jevDecisionPointKo":"추출한 요소와 동작 후보에서 다음 단계를 선택하면 실행기가 데스크톱을 조작한다.","highlightBenefitJa":"画面の読み取り、操作選択、文章生成を分ける。","highlightBenefitKo":"화면 읽기, 동작 선택, 텍스트 생성을 분리한다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"confirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/awlevin/typesafe-computer-use/blob/cc7b5066ae1a07b5e3182e8f87a9b5b6dfdcffc1/typesafe_computer_use/decide.py","note":"Fixed-commit source reviewed / 已核对固定版本源码"},{"url":"https://github.com/awlevin/typesafe-computer-use/blob/cc7b5066ae1a07b5e3182e8f87a9b5b6dfdcffc1/README.md","note":"Fixed-commit source reviewed / 已核对固定版本源码"}],"pinned":false,"topics":["ai-agents","automation","computer-use","macos","ocr","typesafe"]},{"id":"fhshaik:typesafe-mario","name":"typesafe-mario","author":"fhshaik","url":"https://github.com/fhshaik/typesafe-mario","category":"High-Frequency & Simulation","plainSummary":"从 NES 模拟器 RAM 和状态数据中提取环境，让 Jev 选择超级马力欧的手柄按键。","plainSummaryEn":"An experimental NES Mario controller that gives Jev structured emulator RAM and telemetry instead of screenshots.","jevDecisionPoint":"读取运动、敌人、地形及近期操作状态，从预设合法操作中选一项。","jevDecisionPointEn":"Reads motion, enemies, terrain and recent controls, then selects a predefined legal action.","highlightBenefit":"不用给模型发截图，也能记录每次输入状态和选择结果。","highlightBenefitEn":"Records model state and actions without screenshot input.","tags":["games-simulation","typed-decisions"],"stars":332,"forks":34,"openIssues":2,"license":null,"language":"Python","lastCommitAt":"2026-09-16T04:59:07Z","createdAt":"2026-09-16T04:57:42Z","metadataFetchedAt":"2026-09-21T22:02:08.951Z","avatarUrl":"https://avatars.githubusercontent.com/u/123507582?v=4","summarySource":"human-reviewed","claimStatus":"实验控制器；仓库不含游戏 ROM，本站未测持续帧率、完整通关或长程可靠性。","claimStatusEn":"An experimental controller without game ROMs. Sustained rates, full completion and long-horizon reliability were not tested here.","plainSummaryJa":"スクリーンショットではなく NES の RAM と状態を Jev に渡す Mario 操作実験。","plainSummaryKo":"스크린샷 대신 NES RAM과 상태를 Jev에 전달하는 Mario 조작 실험.","jevDecisionPointJa":"動き・敵・地形・直近の操作から、定義済みの合法操作を選ぶ。","jevDecisionPointKo":"움직임·적·지형·최근 조작을 읽고 미리 정의한 합법 동작을 고른다.","highlightBenefitJa":"画像入力なしで、モデルに渡した状態と操作を記録できる。","highlightBenefitKo":"이미지 입력 없이 모델 상태와 동작을 기록한다.","claimStatusJa":"固定バージョンの文書とソースを確認。実行・性能・安全性は当サイトで独立検証していません。","claimStatusKo":"고정 버전 문서와 소스를 확인했습니다. 실행·성능·안전성은 이 사이트에서 독립적으로 검증하지 않았습니다.","licenseStatus":"unconfirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/fhshaik/typesafe-mario/blob/ca22449ed187118d19326d1f54b01b6636578aa4/README.md","note":"README.md"},{"url":"https://github.com/fhshaik/typesafe-mario/blob/ca22449ed187118d19326d1f54b01b6636578aa4/src/typesafe_mario/cli.py","note":"src/typesafe_mario/cli.py"},{"url":"https://github.com/fhshaik/typesafe-mario/blob/ca22449ed187118d19326d1f54b01b6636578aa4/src/typesafe_mario/policy.py","note":"src/typesafe_mario/policy.py"}],"pinned":false},{"id":"RomanSlack:jev-drone","name":"jev-drone","author":"RomanSlack","url":"https://github.com/RomanSlack/jev-drone","category":"High-Frequency & Simulation","plainSummary":"MuJoCo 无人机仿真实验：从相机缓冲区提取场景，Jev 提供战术动作建议。","plainSummaryEn":"A MuJoCo drone experiment deriving scene features from camera buffers for Jev tactical advice.","jevDecisionPoint":"根据距离扇区、障碍高度与目标状态选择机动、评估风险和目标丢失状态。","jevDecisionPointEn":"Uses range sectors, obstacle heights and target state to choose maneuvers and assess risk or target loss.","highlightBenefit":"把战术判断与本地制导、避障和飞控分层。","highlightBenefitEn":"Separates tactical judgments from local guidance, reflexes and flight control.","tags":["games-simulation","typed-decisions"],"stars":111,"forks":10,"openIssues":2,"license":"MIT","language":"Python","lastCommitAt":"2026-09-21T16:51:45Z","createdAt":"2026-09-16T20:54:38Z","metadataFetchedAt":"2026-09-21T22:02:12.007Z","avatarUrl":"https://avatars.githubusercontent.com/u/93448780?v=4","summarySource":"human-reviewed","claimStatus":"2.5/50/500 Hz 是作者架构与仿真设置，不能写成飞行安全保证；未在真实飞行器上核验。 已核对固定版本 README 与集成源码。","claimStatusEn":"The 2.5/50/500 Hz figures describe the author’s simulation architecture. They are not a flight-safety guarantee or validation on a real aircraft. README and integration source reviewed at a fixed commit; not independently run or benchmarked by this site.","plainSummaryJa":"カメラのバッファから場面を抽出し、Jev が戦術を助言する MuJoCo ドローン実験。","plainSummaryKo":"카메라 버퍼에서 장면을 추출하고 Jev가 전술을 조언하는 MuJoCo 드론 실험이다.","jevDecisionPointJa":"距離区分、障害物の高さ、目標状態から機動、リスク、目標喪失を判断する。","jevDecisionPointKo":"거리 구간, 장애물 높이와 목표 상태로 기동, 위험 및 목표 상실을 판단한다.","highlightBenefitJa":"戦術判断をローカルの誘導、反射制御、飛行制御と分ける。","highlightBenefitKo":"전술 판단과 로컬 유도·반사·비행 제어를 분리한다.","claimStatusJa":"固定版の README と実装を確認済み。このサイトでは独立した実行・性能検証をしていません。","claimStatusKo":"고정 버전 README와 구현을 확인했습니다. 이 사이트에서 독립적인 실행·성능 검증은 하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/RomanSlack/jev-drone/blob/cbeb53ce4f17a06ea490ae43effcdad231143610/tactics.py#L184","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/RomanSlack/jev-drone/blob/cbeb53ce4f17a06ea490ae43effcdad231143610/README.md","note":"固定版本的实现与说明 / Pinned source and documentation"}],"pinned":false,"topics":["drone","robotics","mujoco","typesafe","autonomous"]},{"id":"phyous:tsai-sc","name":"tsai-sc","author":"phyous","url":"https://github.com/phyous/tsai-sc","category":"High-Frequency & Simulation","plainSummary":"让 Jev 操作原版 StarCraft shareware 的 Strongarm 关卡，读取状态与推理时暂停游戏。","plainSummaryEn":"A Jev harness for StarCraft shareware’s Strongarm mission, pausing the game during state reads and inference.","jevDecisionPoint":"从结构化游戏状态中选择命令，再通过鼠标和键盘输入执行。","jevDecisionPointEn":"Selects commands from structured game state and executes them with mouse and keyboard input.","highlightBenefit":"作者提供关卡胜利录像与校验报告；这是有限任务实验，不是实时竞技基准。","highlightBenefitEn":"The author provides a victory recording and verification report; this is a bounded mission experiment, not a real-time competitive benchmark.","tags":["games-simulation","typed-decisions"],"stars":21,"forks":1,"openIssues":1,"license":"MIT","language":"Python","lastCommitAt":"2026-09-16T06:37:48Z","createdAt":"2026-09-16T03:17:13Z","metadataFetchedAt":"2026-09-21T22:02:14.958Z","avatarUrl":"https://avatars.githubusercontent.com/u/716386?v=4","summarySource":"human-reviewed","claimStatus":"421 次决策和胜局来自作者的一次指定运行，游戏在推理时暂停。已核对源码与报告。","claimStatusEn":"The 421 decisions and victory are from an author-reported run with the game paused during inference; this site reviewed source and reports but did not reproduce it.","plainSummaryJa":"状態取得と推論中にゲームを停止し、StarCraft shareware の Strongarm を Jev で操作する。","plainSummaryKo":"상태 읽기와 추론 중 게임을 일시 정지하며 Jev로 StarCraft shareware의 Strongarm을 조작한다.","jevDecisionPointJa":"構造化されたゲーム状態からコマンドを選び、マウスとキーボードで実行する。","jevDecisionPointKo":"구조화된 게임 상태에서 명령을 골라 마우스·키보드 입력으로 실행한다.","highlightBenefitJa":"作者による勝利録画と検証報告があるが、限定ミッションの実験でリアルタイム競技の基準ではない。","highlightBenefitKo":"저자의 승리 영상과 검증 보고가 있지만 제한된 임무 실험이며 실시간 경쟁 벤치마크는 아니다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/phyous/tsai-sc/blob/6046ecc60156c4a3c04d384b41821a4ff08501b7/README.md","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/phyous/tsai-sc/blob/6046ecc60156c4a3c04d384b41821a4ff08501b7/tsai_sc/typesafe.py","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/phyous/tsai-sc/blob/6046ecc60156c4a3c04d384b41821a4ff08501b7/tsai_sc/controller.py","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/phyous/tsai-sc/blob/6046ecc60156c4a3c04d384b41821a4ff08501b7/tsai_sc/verify.py","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/phyous/tsai-sc/blob/6046ecc60156c4a3c04d384b41821a4ff08501b7/docs/verified-run.json","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/phyous/tsai-sc/blob/6046ecc60156c4a3c04d384b41821a4ff08501b7/docs/run-environment.json","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/phyous/tsai-sc/blob/6046ecc60156c4a3c04d384b41821a4ff08501b7/docs/visual-review.json","note":"固定版本说明与源码；本站未执行该项目。"}],"pinned":false},{"id":"OpeOginni:oc-plugins","name":"oc-plugins","author":"OpeOginni","url":"https://github.com/OpeOginni/oc-plugins","category":"Security & Guardrails","plainSummary":"OpenCode 插件集合中的 oc-auto-perms，用 Jev 按自然语言规则检查工具操作意图。","plainSummaryEn":"The oc-auto-perms plugin in an OpenCode plugin collection uses Jev to check tool intent against natural-language rules.","jevDecisionPoint":"将拟执行操作与权限规则交给 Jev，再由本地逻辑选择允许、拒绝或询问用户。","jevDecisionPointEn":"Sends proposed actions and permission rules to Jev, then uses local logic to allow, deny, or ask the user.","highlightBenefit":"让权限规则与每次工具操作的判断对应起来。","highlightBenefitEn":"Connects permission policies to judgments about individual tool actions.","tags":["security-guardrails","coding-agents"],"stars":3,"forks":0,"openIssues":0,"license":null,"language":"TypeScript","lastCommitAt":"2026-09-21T16:55:48Z","createdAt":"2026-09-09T13:38:00Z","metadataFetchedAt":"2026-09-21T22:02:18.044Z","avatarUrl":"https://avatars.githubusercontent.com/u/107570612?v=4","summarySource":"human-reviewed","claimStatus":"核实范围为 oc-auto-perms，其他插件不应一起算作 Jev 项目；模型权限检查不是完整安全边界，本站未做攻防测试。","claimStatusEn":"The verified scope is oc-auto-perms, not every plugin in the monorepo. Model-based permission checks are not a complete security boundary; this site has not performed adversarial tests.","plainSummaryJa":"OpenCode プラグイン集の oc-auto-perms が、自然言語ルールに沿ってツール操作の意図を Jev で確認する。","plainSummaryKo":"OpenCode 플러그인 모음의 oc-auto-perms가 자연어 규칙에 따라 도구 작업 의도를 Jev로 확인한다.","jevDecisionPointJa":"操作案と権限ルールを送り、ローカルロジックが許可、拒否、確認を選ぶ。","jevDecisionPointKo":"작업 계획과 권한 규칙을 보내고 로컬 로직이 허용, 거부, 사용자 확인을 선택한다.","highlightBenefitJa":"権限ルールと各操作の判断を対応付ける。","highlightBenefitKo":"권한 정책과 개별 도구 동작의 판단을 연결한다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"unconfirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/OpeOginni/oc-plugins/blob/237f814908a56eb188c33af298c61531e72dbfe0/packages/oc-auto-perms/src/index.ts","note":"Fixed-commit source reviewed / 已核对固定版本源码"},{"url":"https://github.com/OpeOginni/oc-plugins/blob/237f814908a56eb188c33af298c61531e72dbfe0/README.md","note":"Fixed-commit source reviewed / 已核对固定版本源码"}],"pinned":false},{"id":"ufec:jev-block-android-ad","name":"jev-block-android-ad","author":"ufec","url":"https://github.com/ufec/jev-block-android-ad","category":"Security & Guardrails","plainSummary":"Android 通知与短信过滤实验：先执行本地验证码等规则，再让 Jev 判断消息是否是广告噪声。","plainSummaryEn":"An Android notification and SMS filter that applies local OTP rules before asking Jev whether a message is advertising noise.","jevDecisionPoint":"对通过本地门槛的文本做类别判断，由本地映射决定允许或拦截。","jevDecisionPointEn":"Classifies text that passes local gates; local mapping decides whether to allow or suppress it.","highlightBenefit":"将验证码优先放行与不确定时放行的策略写在代码中，方便检查。","highlightBenefitEn":"Makes OTP handling and fail-open behavior explicit in code.","tags":["desktop-os","classification-ranking"],"stars":5,"forks":0,"openIssues":0,"license":"MIT","language":"Kotlin","lastCommitAt":"2026-09-19T16:44:43Z","createdAt":"2026-09-18T15:54:10Z","metadataFetchedAt":"2026-09-21T22:02:20.996Z","avatarUrl":"https://avatars.githubusercontent.com/u/34962267?v=4","summarySource":"human-reviewed","claimStatus":"本地 OTP 匹配并非永不漏判保证；源码中的门槛不能替代真机误拦截测试。","claimStatusEn":"Local OTP matching is not an infallibility guarantee. Code-level gates do not replace device testing for false blocks.","plainSummaryJa":"ローカルの認証コード規則を先に適用し、Jev で広告ノイズを判定する Android フィルター。","plainSummaryKo":"로컬 인증 코드 규칙을 먼저 적용한 뒤 Jev로 광고성 잡음을 판단하는 Android 필터.","jevDecisionPointJa":"ローカル条件を通った文章を分類し、許可・抑制はコードで決める。","jevDecisionPointKo":"로컬 조건을 통과한 텍스트를 분류하고 코드가 허용·억제를 결정한다.","highlightBenefitJa":"認証コード処理と不確かな場合の許可方針を確認できる。","highlightBenefitKo":"인증 코드 처리와 불확실할 때 허용하는 정책을 확인할 수 있다.","claimStatusJa":"固定バージョンの文書とソースを確認。実行・性能・安全性は当サイトで独立検証していません。","claimStatusKo":"고정 버전 문서와 소스를 확인했습니다. 실행·성능·안전성은 이 사이트에서 독립적으로 검증하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/ufec/jev-block-android-ad/blob/8da22d212a3392468686d21a020eb8ae06afe320/README.md","note":"README.md"},{"url":"https://github.com/ufec/jev-block-android-ad/blob/8da22d212a3392468686d21a020eb8ae06afe320/core/decision/src/commonMain/kotlin/me/ethanxu/jevnoisegate/core/decision/TypeSafeBackend.kt","note":"core/decision/src/commonMain/kotlin/me/ethanxu/jevnoisegate/core/decision/TypeSafeBackend.kt"}],"pinned":false,"topics":["ad-blocker","android","jev","kotlin","notification-blocker","sms-blocker","typesafe-ai"]},{"id":"sac-y:jev-cu","name":"Jev-cu","author":"Sac-Y","url":"https://github.com/Sac-Y/Jev-cu","category":"Browser & OS Action","plainSummary":"Codex Computer Use 辅助循环：界面文字候选交给 Jev，读取与执行交给桌面工具。","plainSummaryEn":"A Codex Computer Use loop sending text candidates to Jev while desktop tools observe and act.","jevDecisionPoint":"选择元素、动作、完成度和风险，本地策略决定执行、停止或要求确认。","jevDecisionPointEn":"Chooses targets and actions and assesses completion and risk; local policy controls execution or confirmation.","highlightBenefit":"以文字候选表达判断输入，默认先 dry-run。","highlightBenefitEn":"Uses text candidates and starts with a dry run by default.","tags":["desktop-os","coding-agents"],"stars":541,"forks":54,"openIssues":6,"license":null,"language":"JavaScript","lastCommitAt":"2026-09-21T07:30:29Z","createdAt":"2026-09-18T16:56:08Z","metadataFetchedAt":"2026-09-21T22:02:23.959Z","avatarUrl":"https://avatars.githubusercontent.com/u/270291518?v=4","summarySource":"human-reviewed","claimStatus":"已核对固定版本 README 与集成源码。","claimStatusEn":"README and integration source reviewed at a fixed commit; not independently run or benchmarked by this site.","plainSummaryJa":"画面の文字候補を Jev に送り、観察と実行をデスクトップツールが担う Codex ループ。","plainSummaryKo":"화면 텍스트 후보를 Jev에 보내고 데스크톱 도구가 관찰·실행하는 Codex 루프다.","jevDecisionPointJa":"対象と操作を選び、完了やリスクを判断し、ローカル規則が実行・確認を決める。","jevDecisionPointKo":"대상·동작·완료·위험을 판단하며 로컬 정책이 실행이나 확인을 결정한다.","highlightBenefitJa":"文字候補を入力にし、既定では dry-run から始める。","highlightBenefitKo":"텍스트 후보를 사용하고 기본적으로 dry-run부터 시작한다.","claimStatusJa":"固定版の README と実装を確認済み。このサイトでは独立した実行・性能検証をしていません。","claimStatusKo":"고정 버전 README와 구현을 확인했습니다. 이 사이트에서 독립적인 실행·성능 검증은 하지 않았습니다.","licenseStatus":"unconfirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/Sac-Y/Jev-cu/blob/38fb31de7dfe6209bbe6e04057c00c6e885ba577/scripts/jev-decide.mjs#L6","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/Sac-Y/Jev-cu/blob/38fb31de7dfe6209bbe6e04057c00c6e885ba577/README.md","note":"固定版本的实现与说明 / Pinned source and documentation"}],"pinned":false,"topics":["codex","computer-use","jev","system-one","typesafe"]},{"id":"tamaratran:jev-pruner","name":"jev-pruner","author":"tamaratran","url":"https://github.com/tamaratran/jev-pruner","category":"Context GC & Filter","plainSummary":"Claude Code 输出修剪插件，在 Bash 执行后、结果进入主模型前筛掉部分冗余文本。","plainSummaryEn":"A Claude Code output-pruning plugin filtering selected Bash output before it reaches the main model.","jevDecisionPoint":"满足长度与内容条件后，Jev 按块判断哪些输出应保留，原始内容另行归档。","jevDecisionPointEn":"After length and content checks, Jev judges which chunks to retain while original output is archived.","highlightBenefit":"较短输出、错误和识别出的结构化或源码内容直接保留，避免把所有日志一刀切。","highlightBenefitEn":"Short output, errors and recognized structured or source content pass through unchanged.","tags":["context-compaction","coding-agents"],"stars":133,"forks":9,"openIssues":15,"license":"MIT","language":"TypeScript","lastCommitAt":"2026-09-19T10:18:30Z","createdAt":"2026-09-18T03:00:58Z","metadataFetchedAt":"2026-09-21T22:02:27.018Z","avatarUrl":"https://avatars.githubusercontent.com/u/65584429?v=4","summarySource":"human-reviewed","claimStatus":"已核对固定版本 README 与调用源码。","claimStatusEn":"Reviewed the fixed-version README and call path.","plainSummaryJa":"Bash 実行後、主モデルへ渡す前に一部の出力を絞る Claude Code プラグイン。","plainSummaryKo":"Bash 실행 후 주 모델에 전달하기 전 일부 출력을 줄이는 Claude Code 플러그인.","jevDecisionPointJa":"長さと内容を確認してから Jev が保持する塊を選び、原文は別途保存する。","jevDecisionPointKo":"길이·내용 조건 확인 후 Jev가 유지할 청크를 선택하고 원본은 별도 보관한다.","highlightBenefitJa":"短い出力・エラー・認識された構造化データやソースは変更せず通す。","highlightBenefitKo":"짧은 출력·오류·인식된 구조화 데이터와 소스는 그대로 통과시킨다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/tamaratran/jev-pruner/blob/c5bf7df482b9f8df3dcb838034f06da0d8d0d07a/README.md","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/tamaratran/jev-pruner/blob/c5bf7df482b9f8df3dcb838034f06da0d8d0d07a/src/jev.ts","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/tamaratran/jev-pruner/blob/c5bf7df482b9f8df3dcb838034f06da0d8d0d07a/src/index.ts","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/tamaratran/jev-pruner/blob/c5bf7df482b9f8df3dcb838034f06da0d8d0d07a/src/output.ts","note":"固定版本说明与源码；本站未执行该项目。"}],"pinned":false},{"id":"realzachi:typesafe-adblock","name":"typesafe-adblock","author":"realZachi","url":"https://github.com/realZachi/typesafe-adblock","category":"Browser & OS Action","plainSummary":"一个实验性 Chrome 扩展，让 Jev 判断候选 DOM 元素是否是广告，再高亮或移除。","plainSummaryEn":"An experimental Chrome extension that asks Jev whether candidate DOM elements are ads, then highlights or removes them.","jevDecisionPoint":"对元素文本、标签和链接等信息批量询问 Noul，按阈值应用页面操作。","jevDecisionPointEn":"Batches element text, labels, and link information into Noul questions and applies a threshold.","highlightBenefit":"展示语义判断如何连接到具体页面元素。","highlightBenefitEn":"Shows how semantic judgments can be connected to concrete page elements.","tags":["browser-automation","classification-ranking"],"stars":67,"forks":4,"openIssues":1,"license":"MIT","language":"JavaScript","lastCommitAt":"2026-09-17T17:34:33Z","createdAt":"2026-09-17T17:35:21Z","metadataFetchedAt":"2026-09-21T22:02:29.975Z","avatarUrl":"https://avatars.githubusercontent.com/u/109552121?v=4","summarySource":"human-reviewed","claimStatus":"作者明确称趣味实验：会漏拦与误删，不处理跟踪、恶意软件或视频广告；需要自带 API key。","claimStatusEn":"The author calls it a fun experiment: it can miss ads or remove non-ads and does not handle tracking, malware, or video ads. Requires the user’s API key.","plainSummaryJa":"候補 DOM が広告かを Jev に尋ね、強調表示や削除を行う実験的 Chrome 拡張。","plainSummaryKo":"후보 DOM의 광고 여부를 Jev에 물어 강조하거나 제거하는 실험적 Chrome 확장.","jevDecisionPointJa":"要素の文字、ラベル、リンク情報を Noul で評価し、閾値を適用する。","jevDecisionPointKo":"요소 텍스트, 라벨, 링크 정보를 Noul로 평가하고 임계값을 적용한다.","highlightBenefitJa":"意味的な判断とページ要素の操作を結び付ける例。","highlightBenefitKo":"의미 기반 판단을 구체적인 페이지 요소와 연결하는 예제.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"confirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/realZachi/typesafe-adblock/blob/7e067d243d87b7fe4d511653c0ddcd77b9beee18/src/typesafe.js","note":"Fixed-commit source reviewed / 已核对固定版本源码"},{"url":"https://github.com/realZachi/typesafe-adblock/blob/7e067d243d87b7fe4d511653c0ddcd77b9beee18/README.md","note":"Fixed-commit source reviewed / 已核对固定版本源码"}],"pinned":false},{"id":"red5d:jev-cvss","name":"jev-cvss","author":"Red5d","url":"https://github.com/Red5d/jev-cvss","category":"Security & Guardrails","plainSummary":"用 Jev 从漏洞描述中选择 CVSS 指标，再由 Python 计算 v3.0、v3.1 或 v4.0 分数。","plainSummaryEn":"Scripts that use Jev to select CVSS metrics from vulnerability descriptions, then compute v3.0, v3.1 or v4.0 scores in Python.","jevDecisionPoint":"将攻击条件与影响映射为离散指标；数值计算由本地 CVSS 公式完成。","jevDecisionPointEn":"Maps attack conditions and impacts to discrete metrics; local CVSS formulas calculate the numeric score.","highlightBenefit":"可以检查每个选中的指标及最终向量，而不只看一个总分。","highlightBenefitEn":"Exposes the selected metrics and vector alongside the final score.","tags":["security-guardrails","classification-ranking"],"stars":2,"forks":1,"openIssues":0,"license":"MIT","language":"Python","lastCommitAt":"2026-09-18T16:05:47Z","createdAt":"2026-09-18T15:58:27Z","metadataFetchedAt":"2026-09-21T22:02:32.991Z","avatarUrl":"https://avatars.githubusercontent.com/u/1693880?v=4","summarySource":"human-reviewed","claimStatus":"公式计算可确定，但模型选指标仍可能出错；不是渗透测试或漏洞真实影响验证。","claimStatusEn":"Score formulas are deterministic but model-selected metrics may be wrong; not penetration testing or validation of actual exploit impact. Not run here.","plainSummaryJa":"脆弱性の説明から Jev が CVSS 指標を選び、Python で v3.0・v3.1・v4.0 の値を計算する。","plainSummaryKo":"취약점 설명에서 Jev가 CVSS 지표를 고르고 Python이 v3.0·v3.1·v4.0 점수를 계산한다.","jevDecisionPointJa":"攻撃条件と影響を離散指標へ対応付け、数値はローカルの式で求める。","jevDecisionPointKo":"공격 조건과 영향을 이산 지표에 매핑하고 수치는 로컬 공식으로 구한다.","highlightBenefitJa":"総合点だけでなく、選んだ指標とベクトルを確認できる。","highlightBenefitKo":"최종 점수와 함께 선택 지표와 벡터를 확인할 수 있다.","claimStatusJa":"固定バージョンの文書とソースを確認。実行・性能・安全性は当サイトで独立検証していません。","claimStatusKo":"고정 버전 문서와 소스를 확인했습니다. 실행·성능·안전성은 이 사이트에서 독립적으로 검증하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/Red5d/jev-cvss/blob/b0fdc446e4b9ef070abd570db347a277e496d70c/README.md","note":"README.md"},{"url":"https://github.com/Red5d/jev-cvss/blob/b0fdc446e4b9ef070abd570db347a277e496d70c/cvss3_jev.py","note":"cvss3_jev.py"}],"pinned":false,"topics":["cvss","jev","typesafe-ai"]},{"id":"kavehmz:typesafe-playground","name":"typesafe-playground","author":"kavehmz","url":"https://github.com/kavehmz/typesafe-playground","category":"Evaluation & Observability","plainSummary":"可交互的 Jev 实验集，展示客服工单分流预览和三维驾驶仿真。","plainSummaryEn":"Interactive Jev experiments for support-routing previews and 3D driving simulations.","jevDecisionPoint":"给客服消息做多项判断，或根据结构化模拟传感器选择车道与目标速度。","jevDecisionPointEn":"Judges support messages or chooses lanes and target speed from structured simulated sensors.","highlightBenefit":"把输入、概率和后续行为放在界面中对照观察。","highlightBenefitEn":"Shows inputs, probabilities and resulting behavior together.","tags":["games-simulation","classification-ranking","evaluation-benchmarks"],"stars":11,"forks":3,"openIssues":0,"license":null,"language":"JavaScript","lastCommitAt":"2026-09-21T08:47:03Z","createdAt":"2026-09-17T15:45:03Z","metadataFetchedAt":"2026-09-21T22:02:35.997Z","avatarUrl":"https://avatars.githubusercontent.com/u/230843?v=4","summarySource":"human-reviewed","claimStatus":"不会发送真实工单或退款；Jev 读取结构化仿真传感器，不读取画面，也不是实车自动驾驶验证。 已核对固定版本 README 与集成源码。","claimStatusEn":"The support demo sends no real tickets or refunds. Jev reads structured simulated sensors, not images; this is not real-vehicle autonomous-driving validation. README and integration source reviewed at a fixed commit; not independently run or benchmarked by this site.","plainSummaryJa":"問い合わせ振り分けのプレビューと3D運転シミュレーションを行う Jev 実験集。","plainSummaryKo":"지원 요청 라우팅 미리보기와 3D 주행 시뮬레이션을 제공하는 Jev 실험이다.","jevDecisionPointJa":"問い合わせを評価するか、構造化された模擬センサーから車線と速度を選ぶ。","jevDecisionPointKo":"지원 메시지를 평가하거나 구조화된 모의 센서로 차선과 목표 속도를 고른다.","highlightBenefitJa":"入力、確率、続く動作を画面で比較できる。","highlightBenefitKo":"입력, 확률과 후속 행동을 화면에서 함께 살펴본다.","claimStatusJa":"固定版の README と実装を確認済み。このサイトでは独立した実行・性能検証をしていません。","claimStatusKo":"고정 버전 README와 구현을 확인했습니다. 이 사이트에서 독립적인 실행·성능 검증은 하지 않았습니다.","licenseStatus":"unconfirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/kavehmz/typesafe-playground/blob/ec6102ef693de3c1775af5c844d48713b6d3f4a2/demo01/server.mjs#L67","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/kavehmz/typesafe-playground/blob/ec6102ef693de3c1775af5c844d48713b6d3f4a2/README.md","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/kavehmz/typesafe-playground/blob/ec6102ef693de3c1775af5c844d48713b6d3f4a2/demo01/README.md","note":"固定版本的实现与说明 / Pinned source and documentation"}],"pinned":false},{"id":"obie:ruby_decision_model","name":"ruby_decision_model","author":"obie","url":"https://github.com/obie/ruby_decision_model","category":"SDK & Decision Frameworks","plainSummary":"Ruby 决策模型客户端，可通过 TypeSafe 原生接口或 OpenRouter 调用 Jev。","plainSummaryEn":"A Ruby decision-model client for Jev through TypeSafe’s native API or OpenRouter.","jevDecisionPoint":"用统一 Client 提交状态与类型化问题，解析选择、概率、评分及用量。","jevDecisionPointEn":"A shared Client sends state and typed questions and parses choices, probabilities, scores and usage.","highlightBenefit":"只依赖 Ruby 标准库，可在统一接口下切换 provider。","highlightBenefitEn":"Uses only the Ruby standard library and supports provider selection behind one interface.","tags":["multilanguage-sdk","typed-decisions"],"stars":48,"forks":2,"openIssues":18,"license":"MIT","language":"Ruby","lastCommitAt":"2026-09-18T17:55:06Z","createdAt":"2026-09-18T15:30:56Z","metadataFetchedAt":"2026-09-21T22:02:38.991Z","avatarUrl":"https://avatars.githubusercontent.com/u/3908?v=4","summarySource":"human-reviewed","claimStatus":"已核对固定版本 README 与调用源码。","claimStatusEn":"Reviewed the fixed-version README and call path.","plainSummaryJa":"TypeSafe のネイティブ API または OpenRouter 経由で Jev を呼ぶ Ruby クライアント。","plainSummaryKo":"TypeSafe 기본 API 또는 OpenRouter로 Jev를 호출하는 Ruby 클라이언트.","jevDecisionPointJa":"共通 Client から状態と質問を送り、選択・確率・スコア・使用量を解析する。","jevDecisionPointKo":"공통 Client가 상태와 질문을 보내 선택지·확률·점수·사용량을 파싱한다.","highlightBenefitJa":"Ruby 標準ライブラリだけで動き、同じインターフェースで provider を選べる。","highlightBenefitKo":"Ruby 표준 라이브러리만 사용하며 동일한 인터페이스에서 provider를 선택한다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/obie/ruby_decision_model/blob/f79a890ce4eaa8f83d8220727319ee7b11416e01/README.md","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/obie/ruby_decision_model/blob/f79a890ce4eaa8f83d8220727319ee7b11416e01/lib/ruby_decision_model/providers/typesafe.rb","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/obie/ruby_decision_model/blob/f79a890ce4eaa8f83d8220727319ee7b11416e01/lib/ruby_decision_model/client.rb","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/obie/ruby_decision_model/blob/f79a890ce4eaa8f83d8220727319ee7b11416e01/lib/ruby_decision_model/response.rb","note":"固定版本说明与源码；本站未执行该项目。"}],"pinned":false},{"id":"joshuaeroman:plasmallm","name":"plasmallm","author":"joshuaeroman","url":"https://github.com/joshuaeroman/plasmallm","category":"MCP & Integrations","plainSummary":"KDE Plasma 桌面助手中的 Jev Decisions 适配器，在部件里显示结构化判断结果。","plainSummaryEn":"A Jev Decisions adapter in a KDE Plasma assistant widget for displaying structured judgments.","jevDecisionPoint":"把当前消息转换为决策问题，通过 TypeSafe 或兼容 Decisions 端点取得回答。","jevDecisionPointEn":"Converts the current message into decision questions and receives answers from TypeSafe or a compatible Decisions endpoint.","highlightBenefit":"在已有桌面助手界面里试用决策模型。","highlightBenefitEn":"Makes decision models available in an existing desktop assistant UI.","tags":["desktop-os","typed-decisions"],"stars":28,"forks":5,"openIssues":2,"license":"GPL-2.0","language":"QML","lastCommitAt":"2026-09-18T03:09:44Z","createdAt":"2026-02-17T06:07:08Z","metadataFetchedAt":"2026-09-21T22:02:42.054Z","avatarUrl":"https://avatars.githubusercontent.com/u/105562420?v=4","summarySource":"human-reviewed","claimStatus":"源码适配器说明每条消息独立评估且无对话记忆；未证明 Jev 控制全部桌面操作或指令分流。","claimStatusEn":"The adapter evaluates each message independently without conversation memory. It does not establish that Jev controls all desktop actions or command routing.","plainSummaryJa":"KDE Plasma のアシスタント部品で構造化判断を表示する Jev Decisions アダプター。","plainSummaryKo":"KDE Plasma 어시스턴트 위젯에서 구조화된 판단을 보여주는 Jev Decisions 어댑터.","jevDecisionPointJa":"現在のメッセージを質問に変換し、TypeSafe または互換 Decisions 端点から回答を得る。","jevDecisionPointKo":"현재 메시지를 질문으로 바꿔 TypeSafe나 호환 Decisions 엔드포인트에서 답을 받는다.","highlightBenefitJa":"既存のデスクトップ画面で判断モデルを試せる。","highlightBenefitKo":"기존 데스크톱 어시스턴트 화면에서 판단 모델을 시험할 수 있다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"confirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/joshuaeroman/plasmallm/blob/b7d4975b31722624343d932a2b4a2d2b8a9952ff/package/contents/ui/adapters/decisions.js","note":"Fixed-commit source reviewed / 已核对固定版本源码"},{"url":"https://github.com/joshuaeroman/plasmallm/blob/b7d4975b31722624343d932a2b4a2d2b8a9952ff/README.md","note":"Fixed-commit source reviewed / 已核对固定版本源码"}],"pinned":false,"topics":["ai-assistants","ai-assistants-offline","ai-assisted-development","anthropic","anthropic-claude","desktop-widget","kde-plasma","kde6","llm","ollama","openai","openai-compatible","plasma6","plasmoid"]},{"id":"alp82:goodwatch-monorepo","name":"goodwatch-monorepo","author":"alp82","url":"https://github.com/alp82/goodwatch-monorepo","category":"Evaluation & Observability","plainSummary":"GoodWatch 仓库内的影视特征评分实验，比较 Jev 对情绪、情节等特征的不同问法和批量方式。","plainSummaryEn":"A film-and-TV attribute-scoring experiment inside GoodWatch comparing Jev question designs and batch sizes.","jevDecisionPoint":"对固定影视样本询问特征是否存在或有多明显，记录分数、耗时和 Token。","jevDecisionPointEn":"Asks whether predefined traits are present or how strongly they appear, recording scores, latency and Token usage.","highlightBenefit":"便于对照问题尺度、输入内容与批处理设计，检查特征判断。","highlightBenefitEn":"Compares rating scales, input variants and batching over a frozen sample.","tags":["creative-multimedia","evaluation-benchmarks"],"stars":38,"forks":2,"openIssues":8,"license":"MIT","language":"Python","lastCommitAt":"2026-09-21T17:59:11Z","createdAt":"2023-02-24T18:47:19Z","metadataFetchedAt":"2026-09-21T22:02:44.979Z","avatarUrl":"https://avatars.githubusercontent.com/u/966342?v=4","summarySource":"human-reviewed","claimStatus":"源码明确是独立评分实验，不证明线上推荐使用 Jev，也没有优于向量推荐的已核验依据。本站未重跑。","claimStatusEn":"The source identifies a standalone scoring experiment, not proof of production recommendation use or superiority to vector recommendations; not rerun here.","plainSummaryJa":"GoodWatch 内で、映像作品の特徴に対する Jev の質問設計とバッチ量を比較する実験。","plainSummaryKo":"GoodWatch 안에서 영상 작품 특성에 대한 Jev 질문 설계와 배치 크기를 비교하는 실험.","jevDecisionPointJa":"定義した特徴の有無・強さを質問し、スコア・遅延・Token を記録する。","jevDecisionPointKo":"정의한 특성의 존재·강도를 묻고 점수·지연·Token을 기록한다.","highlightBenefitJa":"固定サンプルで評価尺度・入力条件・バッチ方式を比較できる。","highlightBenefitKo":"고정 표본에서 평가 척도·입력 조건·배치 방식을 비교한다.","claimStatusJa":"固定バージョンの文書とソースを確認。実行・性能・安全性は当サイトで独立検証していません。","claimStatusKo":"고정 버전 문서와 소스를 확인했습니다. 실행·성능·안전성은 이 사이트에서 독립적으로 검증하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/alp82/goodwatch-monorepo/blob/58aa7b3683f82ee4ea58f7da1007ee76400a7115/README.md","note":"README.md"},{"url":"https://github.com/alp82/goodwatch-monorepo/blob/58aa7b3683f82ee4ea58f7da1007ee76400a7115/goodwatch-flows/windmill/f/dna/prototype_comparison/jev_run.py","note":"goodwatch-flows/windmill/f/dna/prototype_comparison/jev_run.py"}],"pinned":false,"topics":["aggregation","ai","discovery","disneyplus","dna","explore","imdb","metacritic","movies","netflix","prime","recommendation-engine","rotten-tomatoes","streaming","tv"]},{"id":"lavallee:mk-jev-fly-brain","name":"mk-jev-fly-brain","author":"lavallee","url":"https://github.com/lavallee/mk-jev-fly-brain","category":"High-Frequency & Simulation","plainSummary":"在 mk.js 格斗游戏中比较果蝇连接组脉冲仿真、Jev 和规则策略。","plainSummaryEn":"Compares a fly-connectome spiking simulation, Jev and rule policies in the mk.js fighting game.","jevDecisionPoint":"Jev 读取格斗状态，从与其他控制器相同的七个动作中选择下一步。","jevDecisionPointEn":"Jev reads game state and chooses among the same seven moves offered to other controllers.","highlightBenefit":"用不同控制组、对局档案和实验说明比较各部分贡献。","highlightBenefitEn":"Uses control variants, match records and experiment notes to compare contributions.","tags":["games-simulation","evaluation-benchmarks"],"stars":2,"forks":1,"openIssues":0,"license":"MIT","language":"JavaScript","lastCommitAt":"2026-09-18T17:50:16Z","createdAt":"2026-09-18T16:04:09Z","metadataFetchedAt":"2026-09-21T22:02:47.978Z","avatarUrl":"https://avatars.githubusercontent.com/u/2786?v=4","summarySource":"human-reviewed","claimStatus":"这是游戏内的神经网络仿真与模型比较，不是活体果蝇，也不证明生物智能或通用实时能力。 已核对固定版本 README 与集成源码。","claimStatusEn":"This is a game-based spiking-network simulation, not a living fly or evidence of general biological intelligence or real-time capability. README and integration source reviewed at a fixed commit; not independently run or benchmarked by this site.","plainSummaryJa":"mk.js 格闘ゲームで、ハエの接続図によるスパイク仮想回路、Jev、規則方式を比較する。","plainSummaryKo":"mk.js 격투 게임에서 초파리 연결망 기반 스파이크 시뮬레이션, Jev와 규칙 정책을 비교한다.","jevDecisionPointJa":"Jev が試合状態を読み、他の制御器と同じ7動作から選ぶ。","jevDecisionPointKo":"Jev가 경기 상태를 읽고 다른 제어기와 같은 7개 동작에서 선택한다.","highlightBenefitJa":"対照条件、試合記録、実験説明で各部分の寄与を比較する。","highlightBenefitKo":"대조 조건, 경기 기록과 실험 설명으로 각 요소의 기여를 비교한다.","claimStatusJa":"固定版の README と実装を確認済み。このサイトでは独立した実行・性能検証をしていません。","claimStatusKo":"고정 버전 README와 구현을 확인했습니다. 이 사이트에서 독립적인 실행·성능 검증은 하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/lavallee/mk-jev-fly-brain/blob/8217937e6e725b0ce3160c6e1ab0183db8b8aa30/jev.py#L19","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/lavallee/mk-jev-fly-brain/blob/8217937e6e725b0ce3160c6e1ab0183db8b8aa30/README.md","note":"固定版本的实现与说明 / Pinned source and documentation"}],"pinned":false},{"id":"tannermidd:specpi","name":"SpecPi","author":"TannerMidd","url":"https://github.com/TannerMidd/SpecPi","category":"SDK & Decision Frameworks","plainSummary":"Pi 编码 Agent 的配置与扩展集合，包含可选 Jev 顾问，用于能力建议和工作流检查。","plainSummaryEn":"A Pi configuration and extension bundle with an optional Jev advisor for capabilities and workflow checks.","jevDecisionPoint":"Jev 评估候选能力、输出和任务状态；扩展按各项开关应用建议或记录结果。","jevDecisionPointEn":"Jev evaluates candidate capabilities, outputs and task state; per-feature switches control applying or recording advice.","highlightBenefit":"顾问调用失败会回到原流程；不是自动保证更优参数或更低成本。","highlightBenefitEn":"Advisor failures return to the original path rather than guaranteeing better parameters or lower costs.","tags":["coding-agents","mcp-integrations","typed-decisions"],"stars":8,"forks":0,"openIssues":1,"license":"MIT","language":"JavaScript","lastCommitAt":"2026-09-20T17:19:02Z","createdAt":"2026-08-28T01:49:53Z","metadataFetchedAt":"2026-09-21T22:02:50.991Z","avatarUrl":"https://avatars.githubusercontent.com/u/2380471?v=4","summarySource":"human-reviewed","claimStatus":"已核对固定版本 README 与调用源码。","claimStatusEn":"Reviewed the fixed-version README and call path.","plainSummaryJa":"能力提案とワークフロー点検用の任意の Jev 顧問を含む Pi の設定・拡張セット。","plainSummaryKo":"기능 제안과 작업 흐름 점검을 위한 선택적 Jev 고문을 포함하는 Pi 설정·확장 모음.","jevDecisionPointJa":"候補機能・出力・状態を Jev が評価し、機能別設定で提案を適用・記録する。","jevDecisionPointKo":"Jev가 기능 후보·출력·상태를 평가하고 기능별 설정에 따라 제안을 적용·기록한다.","highlightBenefitJa":"顧問の失敗時は元の処理へ戻り、性能や費用の改善を保証しない。","highlightBenefitKo":"고문 호출 실패 시 기존 흐름으로 돌아가며 성능이나 비용 개선을 보장하지 않는다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/TannerMidd/SpecPi/blob/55c4c55e9d8d5c2a5ae803d7f4e5ebd5427a45c3/README.md","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/TannerMidd/SpecPi/blob/55c4c55e9d8d5c2a5ae803d7f4e5ebd5427a45c3/extensions/jev-advisor/client.mjs","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/TannerMidd/SpecPi/blob/55c4c55e9d8d5c2a5ae803d7f4e5ebd5427a45c3/extensions/jev-advisor/index.ts","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/TannerMidd/SpecPi/blob/55c4c55e9d8d5c2a5ae803d7f4e5ebd5427a45c3/extensions/jev-advisor/broker.mjs","note":"固定版本说明与源码；本站未执行该项目。"}],"pinned":false,"topics":["agent","agent-skills","agentic","agentic-ai","agents","harness-ai","pi","pidev"]},{"id":"exyze:rift","name":"rift","author":"exYze","url":"https://github.com/exYze/rift","category":"CLI & Pipelines","plainSummary":"Rust 编程终端 Rift 的可选 TypeSafe 决策客户端，给受限判断调用 Jev。","plainSummaryEn":"An optional TypeSafe decision client in the Rust coding terminal Rift for bounded Jev judgments.","jevDecisionPoint":"向 System One 提交状态和类型化问题，解析答案供终端工作流使用。","jevDecisionPointEn":"Sends state and typed questions to System One and parses answers for terminal workflows.","highlightBenefit":"在生成式编程模型之外接入独立的判断接口。","highlightBenefitEn":"Adds a separate decision interface alongside generative coding models.","tags":["coding-agents","cli-git-gates","multilanguage-sdk"],"stars":10,"forks":1,"openIssues":0,"license":"MIT","language":"Rust","lastCommitAt":"2026-09-18T04:02:41Z","createdAt":"2026-06-11T15:02:33Z","metadataFetchedAt":"2026-09-21T22:02:54.091Z","avatarUrl":"https://avatars.githubusercontent.com/u/39066234?v=4","summarySource":"human-reviewed","claimStatus":"Rift 的整体性能基准不能归因于这条 Jev 接入；本站只核实客户端代码，未执行终端工作流。","claimStatusEn":"Rift’s overall performance benchmarks cannot be attributed to this Jev integration. This site reviewed client code but did not execute terminal workflows.","plainSummaryJa":"Rust 製コーディング端末 Rift にある、任意の TypeSafe 判断クライアント。","plainSummaryKo":"Rust 코딩 터미널 Rift에 포함된 선택적 TypeSafe 판단 클라이언트.","jevDecisionPointJa":"状態と型付き質問を System One に送り、端末処理用に回答を解析する。","jevDecisionPointKo":"상태와 타입 질문을 System One에 보내고 터미널 흐름에서 사용할 답을 해석한다.","highlightBenefitJa":"生成型のコーディングモデルとは別に判断インターフェースを追加する。","highlightBenefitKo":"생성형 코딩 모델과 별도로 판단 인터페이스를 추가한다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"confirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/exYze/rift/blob/b446c3fb8ecf0b0e445b16436fd937262956d010/crates/rift-typesafe/src/lib.rs","note":"Fixed-commit source reviewed / 已核对固定版本源码"},{"url":"https://github.com/exYze/rift/blob/b446c3fb8ecf0b0e445b16436fd937262956d010/README.md","note":"Fixed-commit source reviewed / 已核对固定版本源码"}],"pinned":false},{"id":"brainstormity:jev-moderation-bot","name":"Jev-Moderation-Bot","author":"brainstormity","url":"https://github.com/brainstormity/Jev-Moderation-Bot","category":"Security & Guardrails","plainSummary":"一个 Discord 审核机器人，让 Jev 检查垃圾消息和诈骗链接，并按本地规则逐级警告或禁言。","plainSummaryEn":"A Discord moderation bot using Jev to check spam and scam links, with local warning and timeout rules.","jevDecisionPoint":"评估单条消息或成员近期消息的风险，由 Bot 执行删除、提醒及超时处理。","jevDecisionPointEn":"Assesses messages or recent member history; the bot applies deletion, warnings or timeouts.","highlightBenefit":"把判断、处置记录与人工纠正入口放在同一审核流程中。","highlightBenefitEn":"Connects judgments, moderation logs and manual correction controls.","tags":["security-guardrails","mcp-integrations","classification-ranking"],"stars":42,"forks":4,"openIssues":1,"license":null,"language":"Python","lastCommitAt":"2026-09-18T06:48:54Z","createdAt":"2026-09-17T06:19:44Z","metadataFetchedAt":"2026-09-21T22:02:56.991Z","avatarUrl":"https://avatars.githubusercontent.com/u/240602355?v=4","summarySource":"human-reviewed","claimStatus":"模型审核可能误报，聊天记录会被缓存并用于评估；未验证大规模吞吐、安全效果或平台部署。","claimStatusEn":"Model moderation can produce false positives, and message history is cached for evaluation. Throughput, safety effects and deployment were not verified.","plainSummaryJa":"Jev でスパムや詐欺リンクを確認し、ローカル規則で警告・タイムアウトする Discord Bot。","plainSummaryKo":"Jev로 스팸·사기 링크를 확인하고 로컬 규칙으로 경고·타임아웃을 적용하는 Discord Bot.","jevDecisionPointJa":"メッセージや直近履歴のリスクを評価し、Bot が削除・警告などを行う。","jevDecisionPointKo":"메시지나 최근 기록의 위험을 평가하고 Bot이 삭제·경고 등을 처리한다.","highlightBenefitJa":"判断・対応ログ・手動修正を同じ流れで扱える。","highlightBenefitKo":"판단·처리 로그·수동 수정 기능을 같은 흐름에 연결한다.","claimStatusJa":"固定バージョンの文書とソースを確認。実行・性能・安全性は当サイトで独立検証していません。","claimStatusKo":"고정 버전 문서와 소스를 확인했습니다. 실행·성능·안전성은 이 사이트에서 독립적으로 검증하지 않았습니다.","licenseStatus":"unconfirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/brainstormity/Jev-Moderation-Bot/blob/1629ac80bea758883ee7541ffc654c83acfae4b6/README.md","note":"README.md"},{"url":"https://github.com/brainstormity/Jev-Moderation-Bot/blob/1629ac80bea758883ee7541ffc654c83acfae4b6/typesafe/__init__.py","note":"typesafe/__init__.py"}],"pinned":false},{"id":"sumitrevolt:leadgenrationaivoiceagent","name":"leadgenrationaivoiceagent","author":"sumitrevolt","url":"https://github.com/sumitrevolt/leadgenrationaivoiceagent","category":"Domain & Vertical Tools","plainSummary":"营销与语音平台中的 TypeSafe 实验模块，用 Jev 为预设 Agent 角色选择专长标签。","plainSummaryEn":"An experimental TypeSafe module in a marketing and voice platform chooses specialization labels for agent roles.","jevDecisionPoint":"把角色资料与有限专长候选交给 Choice，代码再映射为能力标签。","jevDecisionPointEn":"Sends role data and bounded specialization candidates to Choice, then maps labels to capabilities.","highlightBenefit":"提供一个业务角色分类接入示例。","highlightBenefitEn":"Provides an example of role classification in a business application.","tags":["domain-workflows","classification-ranking"],"stars":1,"forks":0,"openIssues":13,"license":"MIT","language":"Python","lastCommitAt":"2026-09-21T18:01:09Z","createdAt":"2026-01-02T19:06:41Z","metadataFetchedAt":"2026-09-21T22:02:59.990Z","avatarUrl":"https://avatars.githubusercontent.com/u/88972252?v=4","summarySource":"human-reviewed","claimStatus":"源码只核实角色专长分类，未核实逐轮语音意图或 300ms 性能；API 失败有本地默认值，不能当真实模型置信度。 已核对固定版本 README 与集成源码。","claimStatusEn":"The verified integration classifies role specializations. Per-turn voice-intent routing and 300 ms performance were not established. API-failure defaults are not model confidence. README and integration source reviewed at a fixed commit; not independently run or benchmarked by this site.","plainSummaryJa":"マーケティング・音声基盤内の TypeSafe 実験モジュールが、Agent 役割の専門ラベルを選ぶ。","plainSummaryKo":"마케팅·음성 플랫폼의 TypeSafe 실험 모듈이 Agent 역할의 전문 분야 라벨을 고른다.","jevDecisionPointJa":"役割情報と限定候補を Choice に送り、コードが能力ラベルへ対応付ける。","jevDecisionPointKo":"역할 정보와 한정된 후보를 Choice에 보내고 코드가 능력 라벨로 매핑한다.","highlightBenefitJa":"業務アプリでの役割分類の接続例を示す。","highlightBenefitKo":"업무 앱에서 역할 분류를 연결하는 예제를 제공한다.","claimStatusJa":"固定版の README と実装を確認済み。このサイトでは独立した実行・性能検証をしていません。","claimStatusKo":"고정 버전 README와 구현을 확인했습니다. 이 사이트에서 독립적인 실행·성능 검증은 하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/sumitrevolt/leadgenrationaivoiceagent/blob/0ed2da0ea032eb077a196798d1ed5e6aa8602a2b/app/platform/agent_talent_pool.py#L59","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/sumitrevolt/leadgenrationaivoiceagent/blob/0ed2da0ea032eb077a196798d1ed5e6aa8602a2b/README.md","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/sumitrevolt/leadgenrationaivoiceagent/blob/0ed2da0ea032eb077a196798d1ed5e6aa8602a2b/ops/README.md","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/sumitrevolt/leadgenrationaivoiceagent/blob/0ed2da0ea032eb077a196798d1ed5e6aa8602a2b/app/platform/typesafe_integration.py","note":"固定版本的实现与说明 / Pinned source and documentation"}],"pinned":false},{"id":"agentiloop:agent","name":"Agent","author":"AgentiLoop","url":"https://github.com/AgentiLoop/Agent","category":"Security & Guardrails","plainSummary":"原生 macOS Agent 内的可选 Jev 命令风险顾问，配有 TypeSafeKit 客户端。","plainSummaryEn":"An optional Jev command-risk advisor inside a native macOS Agent, with a TypeSafeKit client.","jevDecisionPoint":"对已通过本地规则的 Shell 命令追加破坏性判断，达到配置阈值时拒绝。","jevDecisionPointEn":"Adds a destructive-risk judgment after local shell checks and refuses commands above the configured threshold.","highlightBenefit":"API 故障通常放行并提示，不能把它当作完整的系统安全防护。","highlightBenefitEn":"API failures generally allow the command with a warning, so it is not complete system protection.","tags":["desktop-os","security-guardrails"],"stars":622,"forks":68,"openIssues":1,"license":null,"language":"Swift","lastCommitAt":"2026-09-21T19:31:42Z","createdAt":"2026-03-11T20:15:09Z","metadataFetchedAt":"2026-09-21T22:03:02.967Z","avatarUrl":"https://avatars.githubusercontent.com/u/52664524?v=4","summarySource":"human-reviewed","claimStatus":"Jev 为可选附加风险判断，调用故障通常放行；未进行系统安全验证。","claimStatusEn":"Jev is an optional additional risk judgment and failures generally fail open; no system-security validation was performed.","plainSummaryJa":"TypeSafeKit を備えたネイティブ macOS Agent 内の任意の Jev コマンドリスク顧問。","plainSummaryKo":"TypeSafeKit을 갖춘 네이티브 macOS Agent 내부의 선택적 Jev 명령 위험 고문.","jevDecisionPointJa":"ローカル Shell 検査後に破壊的リスクを評価し、設定閾値以上なら拒否する。","jevDecisionPointKo":"로컬 Shell 검사 후 파괴적 위험을 추가 평가해 설정 임곗값 이상이면 거부한다.","highlightBenefitJa":"API 障害では通常警告付きで通すため、完全なシステム保護ではない。","highlightBenefitKo":"API 장애 시 대체로 경고와 함께 허용하므로 완전한 시스템 보호 수단은 아니다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/AgentiLoop/Agent/blob/078f87ceca1d1190cc73706ac8ec16442e766a27/README.md","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/AgentiLoop/Agent/blob/078f87ceca1d1190cc73706ac8ec16442e766a27/Agent/Services/JevAdvisor.swift","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/AgentiLoop/Agent/blob/078f87ceca1d1190cc73706ac8ec16442e766a27/TypeSafeKit/Sources/TypeSafeMiddleware/JevProvider.swift","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/AgentiLoop/Agent/blob/078f87ceca1d1190cc73706ac8ec16442e766a27/TypeSafeKit/Sources/TypeSafeKit/TypeSafeClient.swift","note":"固定版本说明与源码；本站未执行该项目。"}],"pinned":false,"topics":["accessibility","agentic-framework","agentic-workflow","ai","ai-agent","apple-intelligence","applescript","automation","claude","imessage","javascript","jxa","llm","macos","mcp","ollama","swift","swiftui","xcode","xpc"]},{"id":"captaincore:captaincore","name":"captaincore","author":"CaptainCore","url":"https://github.com/CaptainCore/captaincore","category":"Security & Guardrails","plainSummary":"WordPress 运维工具 CaptainCore 的 Jev 命令，可询问结构化问题，并给恶意代码扫描结果排复核优先级。","plainSummaryEn":"Jev commands in the WordPress toolkit CaptainCore answer structured questions and prioritize malware scanner findings for review.","jevDecisionPoint":"把扫描规则、命中片段与文件上下文交给 Jev，判断真阳性可能性和建议处理方式。","jevDecisionPointEn":"Sends scan rules, matched snippets, and file context to Jev to assess likely true positives and suggested handling.","highlightBenefit":"将规则扫描产生的候选问题整理为供人工复核的队列。","highlightBenefitEn":"Organizes rule-based scan findings into a queue for human review.","tags":["security-guardrails","cli-git-gates","domain-workflows"],"stars":71,"forks":10,"openIssues":0,"license":"MIT","language":"Go","lastCommitAt":"2026-09-20T16:47:43Z","createdAt":"2018-08-09T00:01:19Z","metadataFetchedAt":"2026-09-21T22:03:06.055Z","avatarUrl":"https://avatars.githubusercontent.com/u/38404679?v=4","summarySource":"human-reviewed","claimStatus":"已核实 ask/triage 命令，未找到支持旧文案“插件兼容性预测”的证据；扫描概率不等于恶意代码判定或安全保证。","claimStatusEn":"The ask and triage commands were verified; the old plugin-compatibility prediction claim was unsupported. A scan probability is not a definitive malware verdict or security guarantee.","plainSummaryJa":"WordPress 運用ツール CaptainCore の Jev コマンドで、構造化質問とマルウェア検出結果の優先順位付けを行う。","plainSummaryKo":"WordPress 운영 도구 CaptainCore의 Jev 명령으로 구조화 질문과 악성 코드 탐지 결과 우선순위를 처리한다.","jevDecisionPointJa":"検出ルール、該当部分、ファイルの文脈から、真陽性の可能性と対応候補を評価する。","jevDecisionPointKo":"탐지 규칙, 일치 부분, 파일 맥락으로 실제 문제일 가능성과 대응 후보를 평가한다.","highlightBenefitJa":"ルールベースの検出を人が確認する順番に整理する。","highlightBenefitKo":"규칙 기반 탐지 결과를 사람이 검토할 순서로 정리한다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"confirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/CaptainCore/captaincore/blob/f54e0669350707b2f4f4c67a0130fdd6574bd0b5/cmd/typesafe.go","note":"Fixed-commit source reviewed / 已核对固定版本源码"},{"url":"https://github.com/CaptainCore/captaincore/blob/f54e0669350707b2f4f4c67a0130fdd6574bd0b5/readme.md","note":"Fixed-commit source reviewed / 已核对固定版本源码"}],"pinned":false},{"id":"knuckles92:openwhisper","name":"OpenWhisper","author":"Knuckles92","url":"https://github.com/Knuckles92/OpenWhisper","category":"Voice & Conversation","plainSummary":"语音听写与会议记录应用，可选用 Jev 检查话题变化、面向记录助手的指令和敏感文本。","plainSummaryEn":"A dictation and meeting-notes app with optional Jev checks for topic changes, note-taker instructions and sensitive text.","jevDecisionPoint":"对转录片段做有边界的判断，辅助触发会议检查点、记录操作或远程文本清理前的筛查。","jevDecisionPointEn":"Makes bounded transcript judgments for checkpoints, note-taking actions or screening before remote text cleanup.","highlightBenefit":"把概率判断与本地阈值、开关和失败回退分开，方便检查每项功能。","highlightBenefitEn":"Keeps model probabilities separate from local thresholds, switches and failure fallbacks.","tags":["voice-conversation","security-guardrails","classification-ranking"],"stars":187,"forks":25,"openIssues":0,"license":"MIT","language":"Python","lastCommitAt":"2026-09-20T16:50:03Z","createdAt":"2024-11-25T08:26:12Z","metadataFetchedAt":"2026-09-21T22:03:08.976Z","avatarUrl":"https://avatars.githubusercontent.com/u/130969111?v=4","summarySource":"human-reviewed","claimStatus":"Jev 是可选文本判断模块，不等于会议内容都在本地处理；功能效果和延迟未在本站实测。","claimStatusEn":"Jev is an optional text-judgment module, not proof that all meeting processing stays local. Effects and latency were not tested here.","plainSummaryJa":"文字起こしと会議記録のアプリ。Jev で話題変化・記録助手への指示・機密的な文章を任意に確認する。","plainSummaryKo":"음성 입력·회의 기록 앱으로 Jev가 주제 변화·기록 도우미 지시·민감한 텍스트를 선택적으로 검사한다.","jevDecisionPointJa":"転記した断片を判定し、会議チェックポイント・記録操作・遠隔文章処理前の確認を補助する。","jevDecisionPointKo":"전사 일부를 판단해 회의 체크포인트·기록 동작·원격 텍스트 정리 전 검사를 돕는다.","highlightBenefitJa":"確率判断とローカル閾値・切替・失敗時処理を分けて確認できる。","highlightBenefitKo":"모델 확률과 로컬 임계값·스위치·실패 시 처리를 분리해 확인한다.","claimStatusJa":"固定バージョンの文書とソースを確認。実行・性能・安全性は当サイトで独立検証していません。","claimStatusKo":"고정 버전 문서와 소스를 확인했습니다. 실행·성능·안전성은 이 사이트에서 독립적으로 검증하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/Knuckles92/OpenWhisper/blob/9e83653df183096104769e302a3c907cb277c551/README.md","note":"README.md"},{"url":"https://github.com/Knuckles92/OpenWhisper/blob/9e83653df183096104769e302a3c907cb277c551/services/typesafe.py","note":"services/typesafe.py"},{"url":"https://github.com/Knuckles92/OpenWhisper/blob/9e83653df183096104769e302a3c907cb277c551/docs/typesafe-fast-judgments.md","note":"docs/typesafe-fast-judgments.md"},{"url":"https://github.com/Knuckles92/OpenWhisper/blob/9e83653df183096104769e302a3c907cb277c551/meeting/agent/typesafe_signals.py","note":"meeting/agent/typesafe_signals.py"}],"pinned":false},{"id":"borislemeec:jev","name":"jev","author":"BorisLeMeec","url":"https://github.com/BorisLeMeec/jev","category":"Codebase & Graph Pathfinding","plainSummary":"Go 编写的 Claude Code 插件，用 Jev 查找相关文件、跨文件回答有界问题并处理大段读取。","plainSummaryEn":"A Go Claude Code plugin using Jev to find files, answer bounded questions across code and handle large reads.","jevDecisionPoint":"根据问题筛选与验证代码相关性，减少直接送入 Agent 的整文件内容。","jevDecisionPointEn":"Screens and verifies code relevance before whole files enter the agent context.","highlightBenefit":"返回可定位的文件线索和判断结果。","highlightBenefitEn":"Returns file locations and judgments for focused inspection.","tags":["ast-code-graph","coding-agents","context-compaction"],"stars":13,"forks":0,"openIssues":1,"license":"MIT","language":"Go","lastCommitAt":"2026-09-18T16:15:28Z","createdAt":"2026-09-17T23:07:07Z","metadataFetchedAt":"2026-09-21T22:03:11.998Z","avatarUrl":"https://avatars.githubusercontent.com/u/16005667?v=4","summarySource":"human-reviewed","claimStatus":"不是模型档位路由器；部分效果比较依赖未公开的私有仓库标注集。 已核对固定版本 README 与集成源码。","claimStatusEn":"This is code retrieval, not model-tier routing. Some reported comparisons rely on labeled private repositories that are not published. README and integration source reviewed at a fixed commit; not independently run or benchmarked by this site.","plainSummaryJa":"ファイル検索、コード全体への限定質問、大きな読み取りを Jev で扱う Go 製 Claude Code プラグイン。","plainSummaryKo":"파일 검색, 코드 전반의 한정된 질문과 큰 파일 읽기를 Jev로 처리하는 Go Claude Code 플러그인이다.","jevDecisionPointJa":"ファイル全体を Agent に渡す前に、質問との関連性を選別・確認する。","jevDecisionPointKo":"파일 전체가 Agent 문맥에 들어가기 전에 질문 관련성을 선별·확인한다.","highlightBenefitJa":"確認対象を絞るためのファイル位置と判断を返す。","highlightBenefitKo":"집중 검토를 위한 파일 위치와 판단을 반환한다.","claimStatusJa":"固定版の README と実装を確認済み。このサイトでは独立した実行・性能検証をしていません。","claimStatusKo":"고정 버전 README와 구현을 확인했습니다. 이 사이트에서 독립적인 실행·성능 검증은 하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/BorisLeMeec/jev/blob/e81c1d006b8b23a616486610f311039088521d0c/internal/typesafe/client.go#L23","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/BorisLeMeec/jev/blob/e81c1d006b8b23a616486610f311039088521d0c/README.md","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/BorisLeMeec/jev/blob/e81c1d006b8b23a616486610f311039088521d0c/bench/README.md","note":"固定版本的实现与说明 / Pinned source and documentation"}],"pinned":false},{"id":"michaelliv:runline","name":"runline","author":"Michaelliv","url":"https://github.com/Michaelliv/runline","category":"SDK & Decision Frameworks","plainSummary":"Runline 的 TypeSafe 插件，把 Jev 判断作为 Agent JavaScript 可调用的动作。","plainSummaryEn":"A TypeSafe plugin exposing Jev decisions as callable actions in Runline Agent JavaScript.","jevDecisionPoint":"evaluate、choice、score、noul 等动作提交问题并保留模型答案和用量。","jevDecisionPointEn":"evaluate, choice, score and noul actions submit questions and retain answers and usage.","highlightBenefit":"可在 Runline 的插件流程中组合判断，不会自动检查所有 Shell 命令安全性。","highlightBenefitEn":"Composes decisions with Runline plugins; it does not automatically inspect every shell command for safety.","tags":["multilanguage-sdk","mcp-integrations","typed-decisions"],"stars":163,"forks":13,"openIssues":1,"license":null,"language":"TypeScript","lastCommitAt":"2026-09-19T14:33:50Z","createdAt":"2026-04-13T14:21:15Z","metadataFetchedAt":"2026-09-21T22:03:15.043Z","avatarUrl":"https://avatars.githubusercontent.com/u/13676242?v=4","summarySource":"human-reviewed","claimStatus":"已核对固定版本 README 与调用源码。","claimStatusEn":"Reviewed the fixed-version README and call path.","plainSummaryJa":"Runline の Agent JavaScript から呼べるアクションとして Jev 判断を公開する TypeSafe プラグイン。","plainSummaryKo":"Runline Agent JavaScript에서 호출할 동작으로 Jev 판단을 제공하는 TypeSafe 플러그인.","jevDecisionPointJa":"evaluate・choice・score・noul が質問を送り、回答と使用量を保持する。","jevDecisionPointKo":"evaluate·choice·score·noul이 질문을 보내 응답과 사용량을 보존한다.","highlightBenefitJa":"他のプラグイン処理と組み合わせられるが、全 Shell コマンドの安全性を自動点検するものではない。","highlightBenefitKo":"다른 플러그인과 조합할 수 있지만 모든 Shell 명령의 안전성을 자동 검사하지 않는다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"unconfirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/Michaelliv/runline/blob/6bdddfa82cd95b6b9a07e57fd93a271ae83a0d1b/README.md","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/Michaelliv/runline/blob/6bdddfa82cd95b6b9a07e57fd93a271ae83a0d1b/packages/runline-plugins/typesafe/src/shared.ts","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/Michaelliv/runline/blob/6bdddfa82cd95b6b9a07e57fd93a271ae83a0d1b/packages/runline-plugins/typesafe/src/index.ts","note":"固定版本说明与源码；本站未执行该项目。"}],"pinned":false,"topics":["agent","api","cli","javascript","npm","plugins","quickjs","sandbox","tools","typescript"]},{"id":"kevthetech143:super-jev","name":"super-jev","author":"Kevthetech143","url":"https://github.com/Kevthetech143/super-jev","category":"SDK & Decision Frameworks","plainSummary":"一个 TypeScript 决策执行框架，把证据、Jev 判断、允许的动作和结果记录串起来。","plainSummaryEn":"A TypeScript decision-to-action harness connecting evidence, Jev judgments, permitted actions, and outcome records.","jevDecisionPoint":"提交类型化问题，校验答案后由领域规则选择注册工具，并检查权限与参数。","jevDecisionPointEn":"Submits typed questions, validates answers, and lets domain rules choose registered tools with permission and argument checks.","highlightBenefit":"保留每次判断、工具调用和结果的本地追踪记录。","highlightBenefitEn":"Keeps local traces of judgments, tool calls, and outcomes.","tags":["multilanguage-sdk","typed-decisions"],"stars":8,"forks":2,"openIssues":4,"license":"MIT","language":"Python","lastCommitAt":"2026-09-21T12:28:54Z","createdAt":"2026-09-17T03:56:36Z","metadataFetchedAt":"2026-09-21T22:03:18.003Z","avatarUrl":"https://avatars.githubusercontent.com/u/184859493?v=4","summarySource":"human-reviewed","claimStatus":"实验性框架；默认 demo 使用标注的 fixture，作者小规模 live smoke 不代表通用生产验证。","claimStatusEn":"An experimental harness. Default demos use labeled fixtures; the author’s small live smoke tests do not establish general production readiness. This site has not run it.","plainSummaryJa":"証拠、Jev の判断、許可された操作、結果の記録をつなぐ TypeScript の実行フレームワーク。","plainSummaryKo":"근거, Jev 판단, 허용된 동작, 결과 기록을 연결하는 TypeScript 실행 프레임워크.","jevDecisionPointJa":"型付き質問と回答を検証し、領域ルールが権限や引数を確認して登録ツールを選ぶ。","jevDecisionPointKo":"타입 질문과 답을 검증하고 도메인 규칙이 권한과 인자를 확인해 등록 도구를 선택한다.","highlightBenefitJa":"判断、ツール呼び出し、結果の履歴を手元に残す。","highlightBenefitKo":"판단, 도구 호출, 결과의 로컬 기록을 보존한다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"confirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/Kevthetech143/super-jev/blob/69bb082587459aa2e78245fbb713aac54d53a204/src/jev.ts","note":"Fixed-commit source reviewed / 已核对固定版本源码"},{"url":"https://github.com/Kevthetech143/super-jev/blob/69bb082587459aa2e78245fbb713aac54d53a204/README.md","note":"Fixed-commit source reviewed / 已核对固定版本源码"}],"pinned":false},{"id":"simota:tenbin","name":"tenbin","author":"simota","url":"https://github.com/simota/tenbin","category":"MCP & Integrations","plainSummary":"一套供编程 Agent 设计 Jev 判断流程的文档、MCP server 和 Skill，支持问题检查、批量评估与阈值校准。","plainSummaryEn":"Documentation, an MCP server and a Skill for designing Jev judgments with question linting, batch evaluation and calibration.","jevDecisionPoint":"调用 Choice、Score、Noul 处理样本，并把评估结果用于本地阈值设计。","jevDecisionPointEn":"Runs Choice, Score and Noul over samples and uses measured results to design local thresholds.","highlightBenefit":"让问题设计、样本测量和运行时规则有对应关系。","highlightBenefitEn":"Connects question design, sample evaluation and runtime rules.","tags":["mcp-integrations","evaluation-benchmarks","typed-decisions"],"stars":2,"forks":1,"openIssues":0,"license":"MIT","language":"TypeScript","lastCommitAt":"2026-09-21T01:19:06Z","createdAt":"2026-09-18T06:35:25Z","metadataFetchedAt":"2026-09-21T22:03:21.031Z","avatarUrl":"https://avatars.githubusercontent.com/u/376835?v=4","summarySource":"human-reviewed","claimStatus":"校准需要自己的标注样本，不代表模型输出确定正确。","claimStatusEn":"Calibration requires labeled samples for the workload and does not make judgments infallible; tests were not rerun here.","plainSummaryJa":"質問 lint・バッチ評価・校正で Jev 判断を設計する文書、MCP server、Skill。","plainSummaryKo":"질문 lint·배치 평가·보정으로 Jev 판단을 설계하는 문서·MCP server·Skill.","jevDecisionPointJa":"サンプルを Choice・Score・Noul で評価し、結果からローカル閾値を設計する。","jevDecisionPointKo":"표본에 Choice·Score·Noul을 실행하고 결과로 로컬 임계값을 설계한다.","highlightBenefitJa":"質問設計・測定・実行時規則を対応付けられる。","highlightBenefitKo":"질문 설계·표본 평가·실행 규칙을 연결한다.","claimStatusJa":"固定バージョンの文書とソースを確認。実行・性能・安全性は当サイトで独立検証していません。","claimStatusKo":"고정 버전 문서와 소스를 확인했습니다. 실행·성능·안전성은 이 사이트에서 독립적으로 검증하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/simota/tenbin/blob/80e797986b2334e988bb8c940329a31fe138fbdc/README.md","note":"README.md"},{"url":"https://github.com/simota/tenbin/blob/80e797986b2334e988bb8c940329a31fe138fbdc/tenbin/src/client.ts","note":"tenbin/src/client.ts"}],"pinned":false,"topics":["agent-skill","calibration","classification","claude-code","confidence","jev","llm","mcp","mcp-server","python","system-one","typesafe-ai","typescript"]},{"id":"qddegtya:qualm","name":"qualm","author":"qddegtya","url":"https://github.com/qddegtya/qualm","category":"SDK & Decision Frameworks","plainSummary":"TypeScript 的 Jev 判断封装，把不确定结果作为显式 unsure 分支处理。","plainSummaryEn":"A TypeScript wrapper for Jev decisions with an explicit unsure branch.","jevDecisionPoint":"保留判断概率与选项类型，低置信度时交给调用方提供的回退分支。","jevDecisionPointEn":"Preserves probabilities and option types and routes uncertainty to a caller-supplied fallback.","highlightBenefit":"用类型约束提醒开发者处理不确定性，不保证模型判断一定正确。","highlightBenefitEn":"Uses types to require uncertainty handling without guaranteeing model correctness.","tags":["multilanguage-sdk","typed-decisions"],"stars":1,"forks":0,"openIssues":4,"license":"MIT","language":"TypeScript","lastCommitAt":"2026-09-18T09:15:50Z","createdAt":"2026-09-18T08:11:54Z","metadataFetchedAt":"2026-09-21T22:03:24.036Z","avatarUrl":"https://avatars.githubusercontent.com/u/773248?v=4","summarySource":"human-reviewed","claimStatus":"已核对固定版本 README 与调用源码。","claimStatusEn":"Reviewed the fixed-version README and call path.","plainSummaryJa":"不確実な結果を明示的な unsure 分岐で扱う TypeScript の Jev ラッパー。","plainSummaryKo":"불확실한 결과를 명시적 unsure 분기로 처리하는 TypeScript Jev 래퍼.","jevDecisionPointJa":"確率と選択肢の型を保持し、不確実なら呼び出し側の代替処理へ渡す。","jevDecisionPointKo":"확률과 선택지 타입을 보존하고 불확실하면 호출자가 제공한 대체 분기로 보낸다.","highlightBenefitJa":"型で不確実性への対応を求めるが、モデルの正しさを保証しない。","highlightBenefitKo":"타입으로 불확실성 처리를 요구하지만 모델의 정확성을 보장하지 않는다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/qddegtya/qualm/blob/e1aacf6844f312ba95eb6a1bd26fad15575a7173/README.md","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/qddegtya/qualm/blob/e1aacf6844f312ba95eb6a1bd26fad15575a7173/README.zh-CN.md","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/qddegtya/qualm/blob/e1aacf6844f312ba95eb6a1bd26fad15575a7173/src/provider.ts","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/qddegtya/qualm/blob/e1aacf6844f312ba95eb6a1bd26fad15575a7173/src/client.ts","note":"固定版本说明与源码；本站未执行该项目。"}],"pinned":false,"topics":["ai","classification","decision-making","esm","jev","llm","system-one","type-safety","typesafe-ai","typescript","zero-dependency"]},{"id":"mithrilman:your-signal","name":"your-signal","author":"MithrilMan","url":"https://github.com/MithrilMan/your-signal","category":"Context GC & Filter","plainSummary":"一个自带 key 的 Chrome 扩展，让 Jev 按个人偏好给 X 信息流评分，再调整帖子的显示方式。","plainSummaryEn":"A BYOK Chrome extension using Jev to score X posts against personal preferences and adjust their display.","jevDecisionPoint":"评估帖子相关性、实质内容、实用性与推广倾向，由本地权重和阈值决定高亮、折叠或隐藏。","jevDecisionPointEn":"Scores relevance, substance, usefulness, and promotion; local weights and thresholds control highlighting, collapsing, or hiding.","highlightBenefit":"让个人信息流规则可调整，显示变化可恢复。","highlightBenefitEn":"Makes personal feed rules adjustable and display changes reversible.","tags":["browser-automation","classification-ranking"],"stars":2,"forks":0,"openIssues":0,"license":"MIT","language":"JavaScript","lastCommitAt":"2026-09-18T18:19:06Z","createdAt":"2026-09-18T14:40:23Z","metadataFetchedAt":"2026-09-21T22:03:27.061Z","avatarUrl":"https://avatars.githubusercontent.com/u/5107375?v=4","summarySource":"human-reviewed","claimStatus":"显示决策在本地执行，但待评分的帖子文本会发送到 TypeSafe；未验证过滤准确率或隐私实现。","claimStatusEn":"Display decisions run locally, but post text submitted for scoring is sent to TypeSafe. Filtering accuracy and privacy implementation have not been independently verified.","plainSummaryJa":"自分のキーで Jev に X の投稿を好み別に採点させ、表示を調整する Chrome 拡張。","plainSummaryKo":"자신의 키로 Jev가 X 게시물을 취향별로 평가하고 표시를 조정하는 Chrome 확장.","jevDecisionPointJa":"関連性、内容、実用性、宣伝傾向を評価し、ローカルの重みと閾値で表示を決める。","jevDecisionPointKo":"관련성, 내용, 유용성, 홍보 성향을 평가하고 로컬 가중치와 임계값으로 표시를 결정한다.","highlightBenefitJa":"個人のフィード設定を調整でき、表示変更を元に戻せる。","highlightBenefitKo":"개인 피드 규칙을 조정하고 표시 변경을 되돌릴 수 있다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"confirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/MithrilMan/your-signal/blob/2970c069231b96f094f6fcfe4c11bfcf2da0ecb2/extension/background.js","note":"Fixed-commit source reviewed / 已核对固定版本源码"},{"url":"https://github.com/MithrilMan/your-signal/blob/2970c069231b96f094f6fcfe4c11bfcf2da0ecb2/README.md","note":"Fixed-commit source reviewed / 已核对固定版本源码"}],"pinned":false,"topics":["browser-extension","byok","chrome-extension","feed-filter","jev","manifest-v3","open-source","privacy","typesafe-ai","x-twitter"]},{"id":"akashpriyadarshii:jev-scout","name":"jev-scout","author":"AkashPriyadarshii","url":"https://github.com/AkashPriyadarshii/jev-scout","category":"Data & Search","plainSummary":"用搜索结果建立真实仓库与 Rust crate 候选，再让 Jev 按需求相关性进行评分和选择。","plainSummaryEn":"A repository and Rust crate search tool that asks Jev to score and select retrieved candidates for a request.","jevDecisionPoint":"比较候选信息与用户需求，对适配度和维护迹象打分，并选择候选。","jevDecisionPointEn":"Compares candidate information with the request, scores fit and maintenance signals, and chooses a candidate.","highlightBenefit":"把推荐限制在检索得到的候选中，保留可点击的来源。","highlightBenefitEn":"Keeps recommendations tied to retrieved candidates and source links.","tags":["search-retrieval","classification-ranking","cli-git-gates"],"stars":2,"forks":0,"openIssues":1,"license":"MIT","language":"Rust","lastCommitAt":"2026-09-21T17:42:08Z","createdAt":"2026-09-18T12:20:55Z","metadataFetchedAt":"2026-09-21T22:03:30.002Z","avatarUrl":"https://avatars.githubusercontent.com/u/272530059?v=4","summarySource":"human-reviewed","claimStatus":"真实候选列表不能保证搜索完整、评分正确或项目可靠；“零幻觉”不是本站验证结论。","claimStatusEn":"Retrieved candidates do not guarantee search coverage, correct scoring or project quality; “zero hallucination” is not a verified conclusion here. Not run here.","plainSummaryJa":"検索で得たリポジトリーと Rust crate の候補を、Jev が依頼に合わせて評価・選択する。","plainSummaryKo":"검색한 저장소와 Rust crate 후보를 Jev가 요청에 맞춰 평가·선택한다.","jevDecisionPointJa":"候補情報と依頼を比べ、適合度や保守の兆候を評価する。","jevDecisionPointKo":"후보 정보와 요청을 비교해 적합성과 유지보수 신호를 평가한다.","highlightBenefitJa":"提案を取得済み候補と出典リンクに結び付ける。","highlightBenefitKo":"추천을 검색된 후보와 출처 링크에 연결한다.","claimStatusJa":"固定バージョンの文書とソースを確認。実行・性能・安全性は当サイトで独立検証していません。","claimStatusKo":"고정 버전 문서와 소스를 확인했습니다. 실행·성능·안전성은 이 사이트에서 독립적으로 검증하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/AkashPriyadarshii/jev-scout/blob/acae39260212267288353451c6977aeab7d6af1d/README.md","note":"README.md"},{"url":"https://github.com/AkashPriyadarshii/jev-scout/blob/acae39260212267288353451c6977aeab7d6af1d/src/jev.rs","note":"src/jev.rs"}],"pinned":false,"topics":["ai-agents","claude-code","cli","code-search","coding-agents","crates-io","developer-tools","github-search","jev","mcp-server","package-manager","repository-search","rust","system-one","typesafe-ai","ureq","zero-hallucination"]},{"id":"icohen007:jev-play-ping-pong","name":"jev-play-ping-pong","author":"Icohen007","url":"https://github.com/Icohen007/jev-play-ping-pong","category":"High-Frequency & Simulation","plainSummary":"让 Jev 在浏览器乒乓球游戏中选择发球方向、回球角度和力度。","plainSummaryEn":"Uses Jev to choose serve direction, return angle and pace in a browser table-tennis game.","jevDecisionPoint":"读取球台结构化遥测，通过 Choice 选择接触方向与击球节奏，再由代码执行输入。","jevDecisionPointEn":"Reads structured telemetry and selects contact and pace through Choice before code executes inputs.","highlightBenefit":"保留动作、时延与对局记录，便于复查一次游戏过程。","highlightBenefitEn":"Keeps action, latency and match records for reviewing a run.","tags":["games-simulation","typed-decisions"],"stars":2,"forks":0,"openIssues":0,"license":"MIT","language":"JavaScript","lastCommitAt":"2026-09-17T12:39:50Z","createdAt":"2026-09-17T11:54:54Z","metadataFetchedAt":"2026-09-21T22:03:33.013Z","avatarUrl":"https://avatars.githubusercontent.com/u/37484485?v=4","summarySource":"human-reviewed","claimStatus":"11–0 是作者一次 Club 难度录像；325ms 中位 API 时延及胜利不能推广为胜率或实时性能保证。 已核对固定版本 README 与集成源码。","claimStatusEn":"The 11–0 result is one author-recorded Club-difficulty match. Its 325 ms median API latency and win do not establish a general win rate or latency guarantee. README and integration source reviewed at a fixed commit; not independently run or benchmarked by this site.","plainSummaryJa":"ブラウザの卓球ゲームで、Jev がサーブ方向、返球角度、強さを選ぶ。","plainSummaryKo":"브라우저 탁구 게임에서 Jev가 서브 방향, 리턴 각도와 강도를 고른다.","jevDecisionPointJa":"構造化された状態を読み、Choice で打球を選んでからコードが入力を実行する。","jevDecisionPointKo":"구조화된 상태를 읽고 Choice로 타구를 선택하면 코드가 입력을 실행한다.","highlightBenefitJa":"動作、遅延、試合記録を残し、実行を確認できる。","highlightBenefitKo":"동작, 지연과 경기 기록을 남겨 실행을 검토할 수 있다.","claimStatusJa":"固定版の README と実装を確認済み。このサイトでは独立した実行・性能検証をしていません。","claimStatusKo":"고정 버전 README와 구현을 확인했습니다. 이 사이트에서 독립적인 실행·성능 검증은 하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/Icohen007/jev-play-ping-pong/blob/af8eb3a1f2a14fcebca5accf23a7875b09b30d5e/src/typesafe.mjs#L1","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/Icohen007/jev-play-ping-pong/blob/af8eb3a1f2a14fcebca5accf23a7875b09b30d5e/README.md","note":"固定版本的实现与说明 / Pinned source and documentation"}],"pinned":false,"topics":["ai-agent","browser-automation","computer-use","game-ai","javascript","jev","typesafe-ai"]},{"id":"abdelstark:typesafe-rs","name":"typesafe-rs","author":"AbdelStark","url":"https://github.com/AbdelStark/typesafe-rs","category":"SDK & Decision Frameworks","plainSummary":"Jev 的社区 Rust 客户端，支持异步请求、可选阻塞接口及本地 mock 测试。","plainSummaryEn":"A community Rust Jev client with async requests, an optional blocking interface and local mock testing.","jevDecisionPoint":"向 TypeSafe 提交状态和命名问题，解析 choice、score、noul 答案。","jevDecisionPointEn":"Sends state and named questions to TypeSafe and parses choice, score and noul answers.","highlightBenefit":"提供客户端配置、重试与错误类型，可与异步或同步 Rust 程序组合。","highlightBenefitEn":"Provides client configuration, retries and error types for async or blocking Rust applications.","tags":["multilanguage-sdk","typed-decisions"],"stars":1,"forks":0,"openIssues":0,"license":"MIT","language":"Rust","lastCommitAt":"2026-09-16T12:22:25Z","createdAt":"2026-09-16T08:51:42Z","metadataFetchedAt":"2026-09-21T22:03:35.982Z","avatarUrl":"https://avatars.githubusercontent.com/u/45264458?v=4","summarySource":"human-reviewed","claimStatus":"已核对固定版本 README 与调用源码。","claimStatusEn":"Reviewed the fixed-version README and call path.","plainSummaryJa":"非同期要求、任意の同期インターフェース、ローカル mock テストに対応するコミュニティ Rust クライアント。","plainSummaryKo":"비동기 요청·선택적 동기 인터페이스·로컬 mock 테스트를 지원하는 커뮤니티 Rust 클라이언트.","jevDecisionPointJa":"状態と名前付き質問を TypeSafe に送り、choice・score・noul を解析する。","jevDecisionPointKo":"상태와 이름 있는 질문을 TypeSafe에 보내 choice·score·noul 응답을 파싱한다.","highlightBenefitJa":"クライアント設定・再試行・エラー型を備え、非同期・同期 Rust アプリに組み込める。","highlightBenefitKo":"클라이언트 설정·재시도·오류 타입으로 비동기·동기 Rust 앱에 연동할 수 있다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/AbdelStark/typesafe-rs/blob/8e8b7a2ae1af5bf7d3e5d831e634fdf328b2fbcc/README.md","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/AbdelStark/typesafe-rs/blob/8e8b7a2ae1af5bf7d3e5d831e634fdf328b2fbcc/crates/typesafe-rs/src/client.rs","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/AbdelStark/typesafe-rs/blob/8e8b7a2ae1af5bf7d3e5d831e634fdf328b2fbcc/crates/typesafe-rs/src/transport.rs","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/AbdelStark/typesafe-rs/blob/8e8b7a2ae1af5bf7d3e5d831e634fdf328b2fbcc/crates/typesafe-rs/src/types/response.rs","note":"固定版本说明与源码；本站未执行该项目。"}],"pinned":false,"topics":["machine-learning","typesafe-ai"]},{"id":"chetaslua:jevmeter","name":"jevmeter","author":"ChetasLua","url":"https://github.com/ChetasLua/jevmeter","category":"Creative Tools","plainSummary":"把视频转成带评分仪表的视频成片：Jev 按选定规则给字幕句子评分，再由渲染器叠加显示。","plainSummaryEn":"Creates edited videos with score meters by asking Jev to rate transcript sentences against selected rubrics.","jevDecisionPoint":"逐句读取转录文本，按预设问题和尺度返回分数，供时间轴上的仪表使用。","jevDecisionPointEn":"Scores transcript sentences with preset questions and scales for meters placed on the video timeline.","highlightBenefit":"把句子评分与对应视频片段对齐，便于逐段回看。","highlightBenefitEn":"Aligns sentence scores with the corresponding video segments for review.","tags":["creative-multimedia","classification-ranking"],"stars":80,"forks":5,"openIssues":1,"license":"MIT","language":"Python","lastCommitAt":"2026-09-17T12:35:06Z","createdAt":"2026-09-17T11:39:01Z","metadataFetchedAt":"2026-09-21T22:03:38.982Z","avatarUrl":"https://avatars.githubusercontent.com/u/140589714?v=4","summarySource":"human-reviewed","claimStatus":"字幕评分是模型对文本的判断，不是外部事实核查或读心；作者自报准确率/成本未由本站复测，处理视频也不代表直播硬实时。","claimStatusEn":"Transcript scoring is a model judgment, not external fact-checking or mind-reading. Author-reported accuracy and cost were not independently retested; video processing does not imply hard real-time streaming.","plainSummaryJa":"字幕の各文を Jev が指定基準で採点し、メーターを重ねた動画を出力する。","plainSummaryKo":"자막 문장을 Jev가 지정 기준으로 평가하고 점수 계기를 겹친 영상을 출력한다.","jevDecisionPointJa":"書き起こした文を質問と尺度で採点し、動画のタイムライン上で表示する。","jevDecisionPointKo":"전사 문장을 질문과 척도로 평가해 영상 타임라인에 표시한다.","highlightBenefitJa":"文のスコアを対応する映像と照合できる。","highlightBenefitKo":"문장 점수를 해당 영상 구간과 비교할 수 있다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"confirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/ChetasLua/jevmeter/blob/cbf8e117b5b8835e3294c3a8ee652c7dfa737a9a/jevmeter/score.py","note":"Fixed-commit source reviewed / 已核对固定版本源码"},{"url":"https://github.com/ChetasLua/jevmeter/blob/cbf8e117b5b8835e3294c3a8ee652c7dfa737a9a/README.md","note":"Fixed-commit source reviewed / 已核对固定版本源码"}],"pinned":false,"topics":["debate","jev","typesafe","video-editing","whisper"]},{"id":"wundercorp:loki","name":"loki","author":"wundercorp","url":"https://github.com/wundercorp/loki","category":"Routing & Cost Optimization","plainSummary":"Loki 可选接入 Jev：提供类型化判断工具，并在新会话开始时从当前 gateway 的模型中选择档位。","plainSummaryEn":"Loki optionally adds Jev typed-judgment tools and routes a new session to a model within the selected gateway.","jevDecisionPoint":"评估首个任务需要的能力，由本地策略选择同一 gateway 的模型并保持会话路由。","jevDecisionPointEn":"Assesses the first task’s capability needs; local policy chooses a same-gateway model and keeps the route sticky.","highlightBenefit":"把模型路由作为显式选项，避免把每一轮都当作重新选模型的机会。","highlightBenefitEn":"Makes model routing explicit while retaining the selected route for the session.","tags":["llm-routing-cost","mcp-integrations","typed-decisions"],"stars":26,"forks":6,"openIssues":1,"license":"MIT","language":"Python","lastCommitAt":"2026-09-17T12:27:18Z","createdAt":"2026-09-11T17:31:16Z","metadataFetchedAt":"2026-09-21T22:03:42.047Z","avatarUrl":"https://avatars.githubusercontent.com/u/269550784?v=4","summarySource":"human-reviewed","claimStatus":"README 说明的是 companion 工具与会话路由，不是可保证完成任务的外部裁判；本站未测降本效果。","claimStatusEn":"The README describes companion tools and session routing, not a guaranteed completion referee. Cost effects were not measured here.","plainSummaryJa":"Loki の任意機能として Jev の判断ツールと、同一 gateway 内の会話モデル選択を追加する。","plainSummaryKo":"Loki의 선택 기능으로 Jev 판단 도구와 같은 gateway 안의 세션 모델 선택을 추가한다.","jevDecisionPointJa":"最初のタスクに必要な能力を評価し、選択したモデルを会話中は維持する。","jevDecisionPointKo":"첫 작업에 필요한 능력을 평가하고 선택한 모델을 세션 동안 유지한다.","highlightBenefitJa":"モデル選択を明示し、会話ごとの経路を保てる。","highlightBenefitKo":"모델 선택을 명시하고 세션별 경로를 유지한다.","claimStatusJa":"固定バージョンの文書とソースを確認。実行・性能・安全性は当サイトで独立検証していません。","claimStatusKo":"고정 버전 문서와 소스를 확인했습니다. 실행·성능·안전성은 이 사이트에서 독립적으로 검증하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/wundercorp/loki/blob/5f1aea7c786f5ea8c5f9b9f2db2b791ec326c535/README.md","note":"README.md"},{"url":"https://github.com/wundercorp/loki/blob/5f1aea7c786f5ea8c5f9b9f2db2b791ec326c535/agent/typesafe_client.py","note":"agent/typesafe_client.py"},{"url":"https://github.com/wundercorp/loki/blob/5f1aea7c786f5ea8c5f9b9f2db2b791ec326c535/agent/jev_auto_router.py","note":"agent/jev_auto_router.py"}],"pinned":false,"topics":["agent","agent-orchestration","agent-skills","agentic-ai","agents","ai","autonomous-agents","cli","developer-tools","hermes-agent","jev","llm","llm-agents","llm-inference","loki","mcp","nous-research","systemone","tool-use","typesafe"]},{"id":"danrwilloughby:snifftest","name":"snifftest","author":"DanRWilloughby","url":"https://github.com/DanRWilloughby/snifftest","category":"Creative Tools","plainSummary":"Markdown 与纯文本写作检查器，用本地规则和可选 Jev 判断标记文风问题。","plainSummaryEn":"A Markdown and plain-text prose linter using local rules and optional Jev judgments.","jevDecisionPoint":"逐段判断重复结尾、套话、过度保留等语义规则，返回每条规则的概率。","jevDecisionPointEn":"Assesses paragraphs against style rules such as redundant closings, clichés and excessive hedging.","highlightBenefit":"报告文件、行号和触发规则，保留原文供作者自行修改。","highlightBenefitEn":"Reports files, lines and rules while leaving revision to the author.","tags":["creative-multimedia","cli-git-gates"],"stars":27,"forks":1,"openIssues":4,"license":"MIT","language":"TypeScript","lastCommitAt":"2026-09-18T15:08:48Z","createdAt":"2026-09-17T22:37:52Z","metadataFetchedAt":"2026-09-21T22:03:45.002Z","avatarUrl":"https://avatars.githubusercontent.com/u/8854211?v=4","summarySource":"human-reviewed","claimStatus":"文风规则命中不证明文本由 AI 写成；这不是可靠的作者身份或 AI 来源检测器。 已核对固定版本 README 与集成源码。","claimStatusEn":"Style-rule hits do not establish that AI wrote a passage. This is not a reliable authorship or AI-origin detector. README and integration source reviewed at a fixed commit; not independently run or benchmarked by this site.","plainSummaryJa":"ローカル規則と任意の Jev 判断で文章を点検する Markdown・テキスト用 linter。","plainSummaryKo":"로컬 규칙과 선택형 Jev 판단으로 문장을 검사하는 Markdown·텍스트 linter다.","jevDecisionPointJa":"冗長な結び、常套句、過度な留保などの文体規則を段落ごとに評価する。","jevDecisionPointKo":"중복 결말, 상투어와 과도한 유보 같은 문체 규칙을 문단별로 평가한다.","highlightBenefitJa":"ファイル、行、規則を示し、修正は書き手に任せる。","highlightBenefitKo":"파일, 줄과 규칙을 표시하며 수정은 작성자에게 맡긴다.","claimStatusJa":"固定版の README と実装を確認済み。このサイトでは独立した実行・性能検証をしていません。","claimStatusKo":"고정 버전 README와 구현을 확인했습니다. 이 사이트에서 독립적인 실행·성능 검증은 하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/DanRWilloughby/snifftest/blob/240653e2b082c114c38fed50a42e7eb311e21219/src/jev.ts#L50","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/DanRWilloughby/snifftest/blob/240653e2b082c114c38fed50a42e7eb311e21219/README.md","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/DanRWilloughby/snifftest/blob/240653e2b082c114c38fed50a42e7eb311e21219/assets/nose/README.md","note":"固定版本的实现与说明 / Pinned source and documentation"}],"pinned":false,"topics":["ai-writing","claude-code","claude-code-plugin","claude-skills","github-action","jev","linter","markdown","pre-commit","prose","typesafe","writing"]},{"id":"albionahoti:refgarden","name":"refgarden","author":"AlbionaHoti","url":"https://github.com/AlbionaHoti/refgarden","category":"Creative Tools","plainSummary":"创意参考图库，汇集 The Met、NASA 与 Cosmos 的素材；本地 Explore 模式可使用 Jev。","plainSummaryEn":"A visual-reference gallery drawing from The Met, NASA and Cosmos, with Jev in local Explore mode.","jevDecisionPoint":"Jev 根据标题和描述挑选搜索词与值得突出的参考素材，不读取图片像素。","jevDecisionPointEn":"Jev selects search phrases and highlights references using titles and descriptions, not image pixels.","highlightBenefit":"保留素材来源链接；公开检索演示不调用 Jev，不能据此证明视觉聚类。","highlightBenefitEn":"Keeps source links; the public search demo does not call Jev and is not evidence of visual clustering.","tags":["creative-multimedia","search-retrieval"],"stars":27,"forks":3,"openIssues":0,"license":"MIT","language":"TypeScript","lastCommitAt":"2026-09-19T23:15:39Z","createdAt":"2026-09-18T10:40:25Z","metadataFetchedAt":"2026-09-21T22:03:48.063Z","avatarUrl":"https://avatars.githubusercontent.com/u/22985657?v=4","summarySource":"human-reviewed","claimStatus":"本地 Explore 使用文字元数据进行 Jev 判断；公开演示不调用 Jev，未验证图片理解或聚类能力。","claimStatusEn":"Local Explore uses text metadata for Jev judgments; the public demo does not call Jev, and image understanding or clustering was not verified.","plainSummaryJa":"The Met・NASA・Cosmos の参考素材を集め、ローカル Explore で Jev を使うギャラリー。","plainSummaryKo":"The Met·NASA·Cosmos 자료를 모으고 로컬 Explore에서 Jev를 사용하는 참고 이미지 갤러리.","jevDecisionPointJa":"Jev は画像の画素ではなくタイトルと説明から検索句や注目素材を選ぶ。","jevDecisionPointKo":"Jev는 이미지 픽셀이 아닌 제목·설명으로 검색어와 강조할 자료를 고른다.","highlightBenefitJa":"出典リンクを保持。公開検索デモは Jev を呼ばず、画像クラスタリングの証拠ではない。","highlightBenefitKo":"출처 링크를 보존하며 공개 검색 데모는 Jev를 호출하지 않아 시각 군집화의 증거가 아니다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/AlbionaHoti/refgarden/blob/6affd25a31dd0ed39fc9f71508732fb7ee0b8acc/README.md","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/AlbionaHoti/refgarden/blob/6affd25a31dd0ed39fc9f71508732fb7ee0b8acc/src/jev-client.ts","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/AlbionaHoti/refgarden/blob/6affd25a31dd0ed39fc9f71508732fb7ee0b8acc/src/decision.ts","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/AlbionaHoti/refgarden/blob/6affd25a31dd0ed39fc9f71508732fb7ee0b8acc/server.ts","note":"固定版本说明与源码；本站未执行该项目。"}],"pinned":false,"topics":["creative-tools","jev","moodboard","typesafe","typescript","visual-search"]},{"id":"shiftynick:jev-axi","name":"jev-axi","author":"shiftynick","url":"https://github.com/shiftynick/jev-axi","category":"CLI & Pipelines","plainSummary":"在命令行调用 Jev 做 pick、rate、check、rank、triage、guard，并可接到 Agent 工具调用前的 hook。","plainSummaryEn":"A CLI for Jev pick, rate, check, rank, triage, and guard tasks, with optional hooks before Agent tool calls.","jevDecisionPoint":"把输入状态和选项转为结构化问题，返回结果或供本地安全策略使用的风险分数。","jevDecisionPointEn":"Converts input state and options into structured questions, returning results or risk scores for local policy.","highlightBenefit":"在脚本和 Agent 工作流里复用同一套判断命令。","highlightBenefitEn":"Reuses the same decision commands in scripts and Agent workflows.","tags":["cli-git-gates","security-guardrails","typed-decisions"],"stars":17,"forks":2,"openIssues":3,"license":"MIT","language":"TypeScript","lastCommitAt":"2026-09-19T12:22:29Z","createdAt":"2026-09-16T23:56:26Z","metadataFetchedAt":"2026-09-21T22:03:50.962Z","avatarUrl":"https://avatars.githubusercontent.com/u/5439863?v=4","summarySource":"human-reviewed","claimStatus":"hook 与风险评分是附加检查，不保证阻止所有危险命令或注入；成本、延迟和防护效果未由本站复测。","claimStatusEn":"Hooks and risk scores are additional checks, not a guarantee against all dangerous commands or injection. Cost, latency, and protection have not been independently retested.","plainSummaryJa":"Jev の pick、rate、check、rank、triage、guard を使う CLI。Agent のツール実行前 hook にも対応する。","plainSummaryKo":"Jev의 pick, rate, check, rank, triage, guard를 사용하는 CLI로 Agent 도구 실행 전 hook도 제공한다.","jevDecisionPointJa":"状態と選択肢を質問に変換し、結果やローカルポリシー用のリスクスコアを返す。","jevDecisionPointKo":"상태와 선택지를 질문으로 바꾸고 결과나 로컬 정책용 위험 점수를 반환한다.","highlightBenefitJa":"スクリプトと Agent で同じ判断コマンドを使える。","highlightBenefitKo":"스크립트와 Agent 흐름에서 같은 판단 명령을 재사용한다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"confirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/shiftynick/jev-axi/blob/db4b6fe56f962463d3924019bd038462a2190b2d/src/client.ts","note":"Fixed-commit source reviewed / 已核对固定版本源码"},{"url":"https://github.com/shiftynick/jev-axi/blob/db4b6fe56f962463d3924019bd038462a2190b2d/README.md","note":"Fixed-commit source reviewed / 已核对固定版本源码"}],"pinned":false,"topics":["agent-safety","ai-agents","axi","claude-code","claude-code-hooks","cli","codex","developer-tools","jev","llm-guardrails","prompt-injection","typesafe"]},{"id":"decrux9812:typesafe-skill-router","name":"typesafe-skill-router","author":"DECRUX9812","url":"https://github.com/DECRUX9812/typesafe-skill-router","category":"Routing & Cost Optimization","plainSummary":"一个默认关闭的 Hermes Agent 插件，先用 Jev 从可用 Skill 中挑出与当前请求相关的一项建议。","plainSummaryEn":"An opt-in Hermes Agent plugin that asks Jev to suggest one relevant skill before a model call.","jevDecisionPoint":"评估请求与技能目录；有合适项时追加建议，没有把握或调用失败时不注入。","jevDecisionPointEn":"Compares the request with the skill roster; inserts a suggestion only when suitable and abstains on failure.","highlightBenefit":"让 Agent 先看到一项有针对性的技能提示，同时保留忽略建议的空间。","highlightBenefitEn":"Adds a targeted skill hint that the Agent can still ignore.","tags":["llm-routing-cost","mcp-integrations"],"stars":8,"forks":2,"openIssues":1,"license":"MIT","language":"Python","lastCommitAt":"2026-09-21T15:09:15Z","createdAt":"2026-09-16T05:14:02Z","metadataFetchedAt":"2026-09-21T22:03:54.038Z","avatarUrl":"https://avatars.githubusercontent.com/u/60716910?v=4","summarySource":"human-reviewed","claimStatus":"推荐不是强制加载或最优性保证；具体费用、Token 节省与首字延迟未在本站测量。","claimStatusEn":"Recommendations do not force loading or guarantee optimality; cost, Token savings and time to first output were not measured here.","plainSummaryJa":"モデル呼び出し前に関連 skill を一つ提案する、任意の Hermes Agent プラグイン。","plainSummaryKo":"모델 호출 전에 관련 skill 하나를 제안하는 선택형 Hermes Agent 플러그인.","jevDecisionPointJa":"依頼と skill 一覧を比較し、適切な候補がある場合だけ提案を挿入する。","jevDecisionPointKo":"요청과 skill 목록을 비교하고 적절한 후보가 있을 때만 제안을 넣는다.","highlightBenefitJa":"Agent が無視する余地を残した、対象を絞る skill ヒントになる。","highlightBenefitKo":"Agent가 무시할 수 있는 여지를 남긴 채 관련 skill을 알려 준다.","claimStatusJa":"固定バージョンの文書とソースを確認。実行・性能・安全性は当サイトで独立検証していません。","claimStatusKo":"고정 버전 문서와 소스를 확인했습니다. 실행·성능·안전성은 이 사이트에서 독립적으로 검증하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/DECRUX9812/typesafe-skill-router/blob/e6cdac26f9ed588b4a94b8a2f7f9f026e1b9faf3/README.md","note":"README.md"},{"url":"https://github.com/DECRUX9812/typesafe-skill-router/blob/e6cdac26f9ed588b4a94b8a2f7f9f026e1b9faf3/typesafe_router/client.py","note":"typesafe_router/client.py"}],"pinned":false,"topics":["ai-agents","hermes-agent","hermes-plugin","jev","llm-agents","plugin-catalog","prompt-caching","routing","skill-routing","typesafe"]},{"id":"jujumilk3:jev-calibration-audit","name":"jev-calibration-audit","author":"jujumilk3","url":"https://github.com/jujumilk3/jev-calibration-audit","category":"Evaluation & Observability","plainSummary":"通过公开 API 和数据测试 Jev 概率校准、选项措辞影响及韩文判断表现。","plainSummaryEn":"Audits Jev calibration, option-wording effects and Korean judgments through public APIs and datasets.","jevDecisionPoint":"收集 Noul 与 Choice 输出，对照标签计算误差、准确率和稳定性。","jevDecisionPointEn":"Collects Noul and Choice outputs and compares them with labels for error, accuracy and stability.","highlightBenefit":"保留逐次调用数据与实验说明，便于检查结论适用范围。","highlightBenefitEn":"Keeps per-call records and experiment notes to qualify conclusions.","tags":["evaluation-benchmarks","classification-ranking"],"stars":0,"forks":0,"openIssues":0,"license":"MIT","language":"Python","lastCommitAt":"2026-09-18T09:31:34Z","createdAt":"2026-09-18T09:31:47Z","metadataFetchedAt":"2026-09-21T22:03:56.981Z","avatarUrl":"https://avatars.githubusercontent.com/u/41659814?v=4","summarySource":"human-reviewed","claimStatus":"校准会随任务、标签与选项变化；作者实验不能视为全领域置信度保证。 已核对固定版本 README 与集成源码。","claimStatusEn":"Calibration varies with tasks, labels and answer options. These experiments do not guarantee reliable confidence in every domain. README and integration source reviewed at a fixed commit; not independently run or benchmarked by this site.","plainSummaryJa":"公開 API とデータで Jev の確率較正、選択肢表現の影響、韓国語判断を調べる。","plainSummaryKo":"공개 API와 데이터로 Jev 확률 보정, 선택지 표현의 영향과 한국어 판단을 조사한다.","jevDecisionPointJa":"Noul と Choice の結果をラベルと比べ、誤差、正確さ、安定性を計算する。","jevDecisionPointKo":"Noul·Choice 결과를 정답과 비교해 오류, 정확도와 안정성을 계산한다.","highlightBenefitJa":"呼び出しごとの記録と実験説明で結論の範囲を確認できる。","highlightBenefitKo":"호출별 기록과 실험 설명으로 결론의 적용 범위를 확인한다.","claimStatusJa":"固定版の README と実装を確認済み。このサイトでは独立した実行・性能検証をしていません。","claimStatusKo":"고정 버전 README와 구현을 확인했습니다. 이 사이트에서 독립적인 실행·성능 검증은 하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/jujumilk3/jev-calibration-audit/blob/daab9e2c2d5d5683bf07f3482deb653c26856219/src/jev_audit/client.py#L12","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/jujumilk3/jev-calibration-audit/blob/daab9e2c2d5d5683bf07f3482deb653c26856219/README.md","note":"固定版本的实现与说明 / Pinned source and documentation"}],"pinned":false},{"id":"olli0103:openclaw-typesafe-ai","name":"openclaw-typesafe-ai","author":"Olli0103","url":"https://github.com/Olli0103/openclaw-typesafe-ai","category":"MCP & Integrations","plainSummary":"OpenClaw 的独立社区插件，只注册一个需显式调用的 typesafe_decide 工具。","plainSummaryEn":"An independent OpenClaw plugin registering one explicitly invoked typesafe_decide tool.","jevDecisionPoint":"把调用者提供的状态和问题发到 TypeSafe，返回 Jev 结构化决策。","jevDecisionPointEn":"Sends caller-provided state and questions to TypeSafe and returns typed Jev decisions.","highlightBenefit":"不注册聊天 provider、自动 hooks 或后台服务，也不是抓取或验证码处理插件。","highlightBenefitEn":"Adds no chat provider, automatic hooks or background service; it is not a scraping or CAPTCHA plugin.","tags":["mcp-integrations","typed-decisions"],"stars":0,"forks":0,"openIssues":0,"license":"MIT","language":"TypeScript","lastCommitAt":"2026-09-18T13:30:19Z","createdAt":"2026-09-18T09:55:08Z","metadataFetchedAt":"2026-09-21T22:04:00.043Z","avatarUrl":"https://avatars.githubusercontent.com/u/51629451?v=4","summarySource":"human-reviewed","claimStatus":"已核对固定版本 README 与调用源码。","claimStatusEn":"Reviewed the fixed-version README and call path.","plainSummaryJa":"明示的に呼ぶ typesafe_decide ツール一つを登録する独立した OpenClaw プラグイン。","plainSummaryKo":"명시적으로 호출하는 typesafe_decide 도구 하나를 등록하는 독립 OpenClaw 플러그인.","jevDecisionPointJa":"呼び出し側の状態と質問を TypeSafe に送り、型付き Jev 判断を返す。","jevDecisionPointKo":"호출자가 제공한 상태·질문을 TypeSafe에 보내 구조화된 Jev 판단을 반환한다.","highlightBenefitJa":"チャット provider・自動 hooks・常駐処理はなく、スクレイピングや CAPTCHA 用ではない。","highlightBenefitKo":"채팅 provider·자동 hooks·백그라운드 서비스가 없으며 스크래핑·CAPTCHA용도 아니다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/Olli0103/openclaw-typesafe-ai/blob/985718448bf38a3a58d18389d376621b9bc54615/README.md","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/Olli0103/openclaw-typesafe-ai/blob/985718448bf38a3a58d18389d376621b9bc54615/src/client.ts","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/Olli0103/openclaw-typesafe-ai/blob/985718448bf38a3a58d18389d376621b9bc54615/src/index.ts","note":"固定版本说明与源码；本站未执行该项目。"}],"pinned":false},{"id":"premo-cloud:typesafe-sdk-java","name":"typesafe-sdk-java","author":"Premo-Cloud","url":"https://github.com/Premo-Cloud/typesafe-sdk-java","category":"SDK & Decision Frameworks","plainSummary":"社区维护的 Java TypeSafe 客户端，并提供 Spring Boot Starter 来配置 Jev 调用。","plainSummaryEn":"A community Java TypeSafe client with a Spring Boot Starter for configuring Jev calls.","jevDecisionPoint":"将 Java 状态与 Noul、Choice、Score 问题发送到 System One，并解析类型化结果。","jevDecisionPointEn":"Sends Java state and Noul, Choice, and Score questions to System One and parses typed results.","highlightBenefit":"在 Java 和 Spring Boot 项目中复用请求、配置与错误处理。","highlightBenefitEn":"Reuses request, configuration, and error-handling code in Java and Spring Boot projects.","tags":["multilanguage-sdk","typed-decisions"],"stars":3,"forks":3,"openIssues":4,"license":"MIT","language":"Java","lastCommitAt":"2026-09-21T14:52:01Z","createdAt":"2026-09-18T12:35:22Z","metadataFetchedAt":"2026-09-21T22:04:03.053Z","avatarUrl":"https://avatars.githubusercontent.com/u/259101885?v=4","summarySource":"human-reviewed","claimStatus":"非官方社区客户端，需要 Java 17+；源码接入已核实，但兼容性和生产运行未由本站测试。","claimStatusEn":"An unofficial community client requiring Java 17+. Integration code was verified, but compatibility and production use were not tested by this site.","plainSummaryJa":"Jev の設定に Spring Boot Starter も使える、コミュニティ製 Java TypeSafe クライアント。","plainSummaryKo":"Jev 설정을 위한 Spring Boot Starter도 제공하는 커뮤니티 Java TypeSafe 클라이언트.","jevDecisionPointJa":"Java の状態と Noul、Choice、Score を送信し、型付きの結果を解析する。","jevDecisionPointKo":"Java 상태와 Noul, Choice, Score를 보내고 타입이 있는 결과를 해석한다.","highlightBenefitJa":"Java と Spring Boot でリクエスト、設定、エラー処理を共用できる。","highlightBenefitKo":"Java와 Spring Boot에서 요청, 설정, 오류 처리 코드를 재사용한다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"confirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/Premo-Cloud/typesafe-sdk-java/blob/833893bc46aa3de2e07352a729ff618ea07b6350/typesafe-sdk/src/main/java/io/github/premocloud/typesafe/TypeSafeClient.java","note":"Fixed-commit source reviewed / 已核对固定版本源码"},{"url":"https://github.com/Premo-Cloud/typesafe-sdk-java/blob/833893bc46aa3de2e07352a729ff618ea07b6350/README.md","note":"Fixed-commit source reviewed / 已核对固定版本源码"}],"pinned":false,"topics":["java","jev","sdk","spring-boot","system-one","system-one-models","typesafe"]},{"id":"2389-research:typesafe-go","name":"typesafe-go","author":"2389-research","url":"https://github.com/2389-research/typesafe-go","category":"SDK & Decision Frameworks","plainSummary":"只依赖 Go 标准库的 TypeSafe System One 客户端，用于提交 Jev 问题并读取结构化答案。","plainSummaryEn":"A TypeSafe System One client that uses only the Go standard library to send Jev questions and read structured answers.","jevDecisionPoint":"把 Go 中定义的状态和 Noul、Choice、Score 问题转换为请求，校验并解析返回值。","jevDecisionPointEn":"Builds requests from state and Noul, Choice or Score questions, then validates and parses responses.","highlightBenefit":"在 Go 程序中复用请求、类型映射和错误处理代码。","highlightBenefitEn":"Reuses transport, typed mappings and error handling in Go applications.","tags":["multilanguage-sdk","typed-decisions"],"stars":2,"forks":0,"openIssues":0,"license":"MIT","language":"Go","lastCommitAt":"2026-09-19T03:16:51Z","createdAt":"2026-09-17T20:46:39Z","metadataFetchedAt":"2026-09-21T22:04:06.044Z","avatarUrl":"https://avatars.githubusercontent.com/u/166545110?v=4","summarySource":"human-reviewed","claimStatus":"无第三方 Go 依赖不等于零供应链风险或服务安全保证。","claimStatusEn":"No third-party Go dependencies does not mean zero supply-chain risk or a service-security guarantee.","plainSummaryJa":"Go 標準ライブラリーだけで Jev の質問送信と構造化回答を扱う TypeSafe クライアント。","plainSummaryKo":"Go 표준 라이브러리만으로 Jev 질문 전송과 구조화된 답을 처리하는 TypeSafe 클라이언트.","jevDecisionPointJa":"状態と Noul・Choice・Score を要求へ変換し、回答を検証・解析する。","jevDecisionPointKo":"상태와 Noul·Choice·Score를 요청으로 바꾸고 응답을 검증·해석한다.","highlightBenefitJa":"Go アプリで通信・型対応・エラー処理を再利用できる。","highlightBenefitKo":"Go 앱에서 통신·타입 매핑·오류 처리를 재사용한다.","claimStatusJa":"固定バージョンの文書とソースを確認。実行・性能・安全性は当サイトで独立検証していません。","claimStatusKo":"고정 버전 문서와 소스를 확인했습니다. 실행·성능·안전성은 이 사이트에서 독립적으로 검증하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/2389-research/typesafe-go/blob/1f9ac50bab4921862327dadc45871bc24b381b3e/README.md","note":"README.md"},{"url":"https://github.com/2389-research/typesafe-go/blob/1f9ac50bab4921862327dadc45871bc24b381b3e/typesafe.go","note":"typesafe.go"}],"pinned":false,"topics":["api-client","go","golang","llm","sdk"]},{"id":"cole-gillespie:typesafe-go","name":"typesafe-go","author":"cole-gillespie","url":"https://github.com/cole-gillespie/typesafe-go","category":"SDK & Decision Frameworks","plainSummary":"非官方 Go SDK，支持类型化答案、重试和 context 取消。","plainSummaryEn":"An unofficial Go SDK with typed answers, retries and context cancellation.","jevDecisionPoint":"把应用状态和问题发送给 Jev，解析 Choice、Score 或 Noul 答案。","jevDecisionPointEn":"Sends application state and questions to Jev and parses Choice, Score or Noul answers.","highlightBenefit":"封装请求和响应解析，便于在应用中接入结构化判断。","highlightBenefitEn":"Packages requests and response parsing for typed decisions in applications.","tags":["multilanguage-sdk","typed-decisions"],"stars":1,"forks":0,"openIssues":0,"license":"MIT","language":"Go","lastCommitAt":"2026-09-17T02:06:23Z","createdAt":"2026-09-17T02:07:24Z","metadataFetchedAt":"2026-09-21T22:04:09.058Z","avatarUrl":"https://avatars.githubusercontent.com/u/745064?v=4","summarySource":"human-reviewed","claimStatus":"已核对固定版本 README 与集成源码。","claimStatusEn":"README and integration source reviewed at a fixed commit; not independently run or benchmarked by this site.","plainSummaryJa":"型付き回答、再試行、context キャンセルを備える非公式 Go SDK。","plainSummaryKo":"타입이 지정된 응답, 재시도와 context 취소를 지원하는 비공식 Go SDK다.","jevDecisionPointJa":"状態と質問を Jev に送り、Choice、Score、Noul の回答を解析する。","jevDecisionPointKo":"상태와 질문을 Jev에 보내 Choice, Score, Noul 답변을 처리한다.","highlightBenefitJa":"リクエストと応答解析をまとめ、アプリに型付き判断を組み込める。","highlightBenefitKo":"요청과 응답 처리를 묶어 앱에 구조화된 판단을 연결한다.","claimStatusJa":"固定版の README と実装を確認済み。このサイトでは独立した実行・性能検証をしていません。","claimStatusKo":"고정 버전 README와 구현을 확인했습니다. 이 사이트에서 독립적인 실행·성능 검증은 하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/cole-gillespie/typesafe-go/blob/188247d639e8aacecd7a0d1aceaf8a4c66f8fa57/client.go#L21","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/cole-gillespie/typesafe-go/blob/188247d639e8aacecd7a0d1aceaf8a4c66f8fa57/README.md","note":"固定版本的实现与说明 / Pinned source and documentation"}],"pinned":false},{"id":"devmortimer:pi-typesafe","name":"pi-typesafe","author":"DevMortimer","url":"https://github.com/DevMortimer/pi-typesafe","category":"MCP & Integrations","plainSummary":"Pi 的 Jev 扩展，提供判断工具、终端试验命令和供其他扩展复用的 API。","plainSummaryEn":"A Pi Jev extension providing a decision tool, terminal playground commands and an API for other extensions.","jevDecisionPoint":"通过共享客户端批量提交判断题，验证响应并记录用量与可用性。","jevDecisionPointEn":"A shared client batches typed questions, validates responses and records usage and availability.","highlightBenefit":"统一密钥与客户端管理，其他扩展可复用同一决策接口。","highlightBenefitEn":"Centralizes key and client handling so other extensions can reuse the decision interface.","tags":["mcp-integrations","coding-agents","typed-decisions"],"stars":35,"forks":3,"openIssues":2,"license":"MIT","language":"TypeScript","lastCommitAt":"2026-09-21T01:59:28Z","createdAt":"2026-09-16T01:35:23Z","metadataFetchedAt":"2026-09-21T22:04:12.028Z","avatarUrl":"https://avatars.githubusercontent.com/u/161127243?v=4","summarySource":"human-reviewed","claimStatus":"已核对固定版本 README 与调用源码。","claimStatusEn":"Reviewed the fixed-version README and call path.","plainSummaryJa":"判断ツール・ターミナル実験コマンド・他拡張向け API を提供する Pi の Jev 拡張。","plainSummaryKo":"판단 도구·터미널 실험 명령·다른 확장용 API를 제공하는 Pi Jev 확장.","jevDecisionPointJa":"共通クライアントが質問を一括送信し、応答を検証して使用量と可用性を記録する。","jevDecisionPointKo":"공통 클라이언트가 질문을 묶어 보내고 응답을 검증하며 사용량과 가용성을 기록한다.","highlightBenefitJa":"キーとクライアントを共通管理し、他拡張でも判断 API を再利用できる。","highlightBenefitKo":"키와 클라이언트를 통합 관리해 다른 확장도 판단 인터페이스를 재사용한다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/DevMortimer/pi-typesafe/blob/b822b8a0e4d729bd0db055a72581f21da686dd42/README.md","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/DevMortimer/pi-typesafe/blob/b822b8a0e4d729bd0db055a72581f21da686dd42/src/client.ts","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/DevMortimer/pi-typesafe/blob/b822b8a0e4d729bd0db055a72581f21da686dd42/src/extension.ts","note":"固定版本说明与源码；本站未执行该项目。"}],"pinned":false},{"id":"marandaneto:typesafe-sdk-swift","name":"typesafe-sdk-swift","author":"marandaneto","url":"https://github.com/marandaneto/typesafe-sdk-swift","category":"SDK & Decision Frameworks","plainSummary":"使用 Swift Package Manager、Swift 并发和 URLSession 调用 TypeSafe 的实验性 Swift SDK。","plainSummaryEn":"An experimental Swift SDK for TypeSafe using Swift Package Manager, Swift concurrency, and URLSession.","jevDecisionPoint":"将状态和类型化问题发送到 System One，以 async/await 获取结构化答案。","jevDecisionPointEn":"Sends state and typed questions to System One and receives structured answers through async/await.","highlightBenefit":"让 Swift 应用复用 Jev 请求与响应处理。","highlightBenefitEn":"Lets Swift applications reuse Jev request and response handling.","tags":["multilanguage-sdk","typed-decisions"],"stars":1,"forks":0,"openIssues":0,"license":"MIT","language":"Swift","lastCommitAt":"2026-09-18T13:47:03Z","createdAt":"2026-09-18T11:42:40Z","metadataFetchedAt":"2026-09-21T22:04:15.088Z","avatarUrl":"https://avatars.githubusercontent.com/u/5731772?v=4","summarySource":"human-reviewed","claimStatus":"作者标注 Experimental；跨平台编译通过不等于各最低系统版本均经过运行测试，本站未测试。","claimStatusEn":"Marked experimental by the author. Cross-platform compilation does not mean runtime testing on every minimum OS version; this site has not tested it.","plainSummaryJa":"Swift Package Manager、Swift 並行処理、URLSession を使う実験的 TypeSafe SDK。","plainSummaryKo":"Swift Package Manager, Swift 동시성, URLSession을 사용하는 실험적 TypeSafe SDK.","jevDecisionPointJa":"状態と型付き質問を System One に送り、async/await で回答を受け取る。","jevDecisionPointKo":"상태와 타입 질문을 System One에 보내고 async/await로 답을 받는다.","highlightBenefitJa":"Swift アプリで Jev のリクエストと応答処理を再利用できる。","highlightBenefitKo":"Swift 애플리케이션에서 Jev 요청과 응답 처리를 재사용한다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"confirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/marandaneto/typesafe-sdk-swift/blob/5a8565791b56d6ba874de9f7ea206b8a1a53c794/Sources/TypeSafe/TypeSafeClient.swift","note":"Fixed-commit source reviewed / 已核对固定版本源码"},{"url":"https://github.com/marandaneto/typesafe-sdk-swift/blob/5a8565791b56d6ba874de9f7ea206b8a1a53c794/README.md","note":"Fixed-commit source reviewed / 已核对固定版本源码"}],"pinned":false},{"id":"codeitlikemiley:typesafe-sdk-rust","name":"typesafe-sdk-rust","author":"codeitlikemiley","url":"https://github.com/codeitlikemiley/typesafe-sdk-rust","category":"SDK & Decision Frameworks","plainSummary":"TypeSafe API 的 Rust 客户端，提供异步与可选阻塞调用，以及带类型的问题和答案封装。","plainSummaryEn":"A Rust client for TypeSafe with asynchronous and optional blocking calls plus typed question and answer wrappers.","jevDecisionPoint":"构造 Jev 请求，处理鉴权、网络调用和结构化响应解析。","jevDecisionPointEn":"Builds Jev requests and handles authentication, transport and structured response parsing.","highlightBenefit":"让 Rust 应用复用接口模型和通信逻辑。","highlightBenefitEn":"Reuses API models and transport logic in Rust applications.","tags":["multilanguage-sdk","typed-decisions"],"stars":3,"forks":0,"openIssues":0,"license":"MIT","language":"Rust","lastCommitAt":"2026-09-17T07:31:17Z","createdAt":"2026-09-16T12:31:37Z","metadataFetchedAt":"2026-09-21T22:04:18.103Z","avatarUrl":"https://avatars.githubusercontent.com/u/28816690?v=4","summarySource":"human-reviewed","claimStatus":"类型封装不能消除所有非法请求或运行时错误。","claimStatusEn":"Type wrappers do not eliminate every invalid request or runtime error; this crate and the live API were not tested here.","plainSummaryJa":"非同期と任意の同期呼び出し、型付き質問・回答を備えた TypeSafe 用 Rust クライアント。","plainSummaryKo":"비동기와 선택형 동기 호출, 타입 질문·답을 제공하는 TypeSafe Rust 클라이언트.","jevDecisionPointJa":"Jev 要求を作り、認証・通信・回答解析を処理する。","jevDecisionPointKo":"Jev 요청을 만들고 인증·통신·응답 해석을 처리한다.","highlightBenefitJa":"Rust アプリで API モデルと通信ロジックを再利用できる。","highlightBenefitKo":"Rust 앱에서 API 모델과 통신 로직을 재사용한다.","claimStatusJa":"固定バージョンの文書とソースを確認。実行・性能・安全性は当サイトで独立検証していません。","claimStatusKo":"고정 버전 문서와 소스를 확인했습니다. 실행·성능·안전성은 이 사이트에서 독립적으로 검증하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/codeitlikemiley/typesafe-sdk-rust/blob/e7050df6baa611bc9088f1a361b4b90e8571929d/README.md","note":"README.md"},{"url":"https://github.com/codeitlikemiley/typesafe-sdk-rust/blob/e7050df6baa611bc9088f1a361b4b90e8571929d/src/client.rs","note":"src/client.rs"}],"pinned":false},{"id":"butochnikov:typesafe-sdk-php","name":"typesafe-sdk-php","author":"Butochnikov","url":"https://github.com/Butochnikov/typesafe-sdk-php","category":"SDK & Decision Frameworks","plainSummary":"面向 PHP 8.2+ 的社区 TypeSafe SDK，提供同步调用与基于 Guzzle 的异步请求。","plainSummaryEn":"A community TypeSafe SDK for PHP 8.2+, with synchronous calls and Guzzle-based async requests.","jevDecisionPoint":"把应用状态和问题发送给 Jev，解析 Choice、Score 或 Noul 答案。","jevDecisionPointEn":"Sends application state and questions to Jev and parses Choice, Score or Noul answers.","highlightBenefit":"封装请求和响应解析，便于在应用中接入结构化判断。","highlightBenefitEn":"Packages requests and response parsing for typed decisions in applications.","tags":["multilanguage-sdk","typed-decisions"],"stars":1,"forks":0,"openIssues":0,"license":"MIT","language":"PHP","lastCommitAt":"2026-09-17T02:48:48Z","createdAt":"2026-09-17T02:42:17Z","metadataFetchedAt":"2026-09-21T22:04:20.991Z","avatarUrl":"https://avatars.githubusercontent.com/u/4212297?v=4","summarySource":"human-reviewed","claimStatus":"已核对固定版本 README 与集成源码。","claimStatusEn":"README and integration source reviewed at a fixed commit; not independently run or benchmarked by this site.","plainSummaryJa":"同期呼び出しと Guzzle 非同期要求を備える PHP 8.2+ 向けコミュニティ TypeSafe SDK。","plainSummaryKo":"동기 호출과 Guzzle 비동기 요청을 제공하는 PHP 8.2+ 커뮤니티 TypeSafe SDK다.","jevDecisionPointJa":"状態と質問を Jev に送り、Choice、Score、Noul の回答を解析する。","jevDecisionPointKo":"상태와 질문을 Jev에 보내 Choice, Score, Noul 답변을 처리한다.","highlightBenefitJa":"リクエストと応答解析をまとめ、アプリに型付き判断を組み込める。","highlightBenefitKo":"요청과 응답 처리를 묶어 앱에 구조화된 판단을 연결한다.","claimStatusJa":"固定版の README と実装を確認済み。このサイトでは独立した実行・性能検証をしていません。","claimStatusKo":"고정 버전 README와 구현을 확인했습니다. 이 사이트에서 독립적인 실행·성능 검증은 하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/Butochnikov/typesafe-sdk-php/blob/d44db82dbf4ff601b123731b517fba1ca9092278/src/TypeSafeClient.php#L32","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/Butochnikov/typesafe-sdk-php/blob/d44db82dbf4ff601b123731b517fba1ca9092278/README.md","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/Butochnikov/typesafe-sdk-php/blob/d44db82dbf4ff601b123731b517fba1ca9092278/README.ru.md","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/Butochnikov/typesafe-sdk-php/blob/d44db82dbf4ff601b123731b517fba1ca9092278/src/Constants.php","note":"固定版本的实现与说明 / Pinned source and documentation"}],"pinned":false},{"id":"typesafe-ai:typesafe-sdk-js","name":"typesafe-sdk-js","author":"typesafe-ai","url":"https://github.com/typesafe-ai/typesafe-sdk-js","category":"SDK & Decision Frameworks","plainSummary":"TypeSafe 组织发布的 JavaScript 与 TypeScript SDK，提供 Jev 请求和回答类型。","plainSummaryEn":"The JavaScript and TypeScript SDK published by TypeSafe, with typed Jev requests and answers.","jevDecisionPoint":"systemOne 提交状态与命名问题，并按问题类型推导返回答案。","jevDecisionPointEn":"systemOne submits state and named questions and infers answer types from the questions.","highlightBenefit":"支持 ESM、CommonJS 和 TypeScript 类型声明，便于直接接入应用。","highlightBenefitEn":"Includes ESM, CommonJS and TypeScript declarations for application integration.","tags":["multilanguage-sdk","typed-decisions"],"stars":212,"forks":21,"openIssues":13,"license":"MIT","language":"TypeScript","lastCommitAt":"2026-09-15T18:16:33Z","createdAt":"2026-09-04T22:16:47Z","metadataFetchedAt":"2026-09-21T22:04:24.107Z","avatarUrl":"https://avatars.githubusercontent.com/u/171090879?v=4","summarySource":"human-reviewed","claimStatus":"已核对固定版本 README 与调用源码。","claimStatusEn":"Reviewed the fixed-version README and call path.","plainSummaryJa":"TypeSafe が公開する JavaScript / TypeScript SDK。Jev の要求と回答に型を提供する。","plainSummaryKo":"TypeSafe가 공개한 JavaScript·TypeScript SDK로 Jev 요청·응답 타입을 제공한다.","jevDecisionPointJa":"systemOne が状態と名前付き質問を送り、質問から回答型を推論する。","jevDecisionPointKo":"systemOne이 상태와 이름 있는 질문을 보내고 질문에서 응답 타입을 추론한다.","highlightBenefitJa":"ESM・CommonJS・TypeScript 型宣言を含み、アプリへ組み込める。","highlightBenefitKo":"ESM·CommonJS·TypeScript 타입 선언을 포함해 앱에 연동할 수 있다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/typesafe-ai/typesafe-sdk-js/blob/66880ccded6cb642dc1809620c2b108c33730214/README.md","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/typesafe-ai/typesafe-sdk-js/blob/66880ccded6cb642dc1809620c2b108c33730214/src/client.ts","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/typesafe-ai/typesafe-sdk-js/blob/66880ccded6cb642dc1809620c2b108c33730214/src/types.ts","note":"固定版本说明与源码；本站未执行该项目。"}],"pinned":false},{"id":"wy-coliney:jev-browser-use","name":"jev-browser-use","author":"wy-coliney","url":"https://github.com/wy-coliney/jev-browser-use","category":"Browser & OS Action","plainSummary":"给 Codex 浏览器工作流加一个 Skill：Jev 选导航、点击和滚动，Codex 保留文字输入与最终核验。","plainSummaryEn":"A Skill for Codex browser workflows: Jev chooses navigation, clicks, and scrolling while Codex handles text input and final checks.","jevDecisionPoint":"将页面状态与可执行候选交给 Jev，现有浏览器连接负责执行动作。","jevDecisionPointEn":"Sends page state and permitted candidates to Jev; the existing browser connection executes the actions.","highlightBenefit":"把重复的页面选择交给独立决策步骤，复用已有浏览器连接。","highlightBenefitEn":"Separates repetitive page choices into decision steps while reusing the browser connection.","tags":["browser-automation","coding-agents"],"stars":319,"forks":26,"openIssues":3,"license":"MIT","language":"JavaScript","lastCommitAt":"2026-09-19T02:14:59Z","createdAt":"2026-09-18T06:58:34Z","metadataFetchedAt":"2026-09-21T22:04:27.011Z","avatarUrl":"https://avatars.githubusercontent.com/u/30455008?v=4","summarySource":"human-reviewed","claimStatus":"需要另行配置 Codex 浏览器连接；作者报告的 5–10 倍针对其工作流，不是通用性能保证。","claimStatusEn":"Requires a separately configured Codex browser connection. The author’s reported 5–10× speedup concerns their own workflows, not a general guarantee; this site has not retested it.","plainSummaryJa":"Codex のブラウザー作業で、Jev が移動、クリック、スクロールを選び、文字入力と最終確認は Codex が行う Skill。","plainSummaryKo":"Codex 브라우저 작업에서 Jev가 탐색, 클릭, 스크롤을 선택하고 입력과 최종 확인은 Codex가 맡는 Skill.","jevDecisionPointJa":"ページ状態と実行候補を Jev に送り、既存のブラウザー接続で操作する。","jevDecisionPointKo":"페이지 상태와 실행 후보를 Jev에 보내고 기존 브라우저 연결로 동작을 실행한다.","highlightBenefitJa":"繰り返す画面選択を独立させ、既存の接続を再利用する。","highlightBenefitKo":"반복되는 화면 선택을 분리하면서 기존 연결을 재사용한다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"confirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/wy-coliney/jev-browser-use/blob/f14b60e0ae1ee90cd73eb6650e30a666a84c021a/skills/jev-browser-use/bridge.mjs","note":"Fixed-commit source reviewed / 已核对固定版本源码"},{"url":"https://github.com/wy-coliney/jev-browser-use/blob/f14b60e0ae1ee90cd73eb6650e30a666a84c021a/README.md","note":"Fixed-commit source reviewed / 已核对固定版本源码"}],"pinned":false},{"id":"bensyverson:goodall","name":"goodall","author":"bensyverson","url":"https://github.com/bensyverson/goodall","category":"SDK & Decision Frameworks","plainSummary":"Go Agent 库中的可选 TypeSafe 包，允许把 Jev 当工具或路由判断使用，而不替代对话模型。","plainSummaryEn":"An optional TypeSafe package in a Go Agent library, using Jev as a tool or routing judge alongside chat models.","jevDecisionPoint":"通过独立客户端提交类型化问题，可用于工具调用、回合路由和邮件分类示例。","jevDecisionPointEn":"Submits typed questions through a separate client for tools, turn routing and a mail-triage example.","highlightBenefit":"把语义判断与生成模型的 Agent 循环分开接入。","highlightBenefitEn":"Keeps typed judgments separate from the generative Agent loop.","tags":["multilanguage-sdk","llm-routing-cost"],"stars":2,"forks":0,"openIssues":0,"license":"MIT","language":"Go","lastCommitAt":"2026-09-18T02:26:29Z","createdAt":"2026-09-15T13:47:08Z","metadataFetchedAt":"2026-09-21T22:04:30.092Z","avatarUrl":"https://avatars.githubusercontent.com/u/2905922?v=4","summarySource":"human-reviewed","claimStatus":"Jev 不是 goodall 的主聊天 provider；邮件示例只输出建议与草稿，不自动发信。","claimStatusEn":"Jev is not goodall’s primary chat provider; the mail example proposes routes and drafts without sending mail. Not run here.","plainSummaryJa":"Go Agent ライブラリー内で、対話モデルとは別に Jev をツールや振り分け判断に使う追加パッケージ。","plainSummaryKo":"Go Agent 라이브러리에서 대화 모델과 별도로 Jev를 도구나 라우팅 판단에 쓰는 선택 패키지.","jevDecisionPointJa":"専用クライアントで質問し、ツール・ターン振り分け・メール分類例に使う。","jevDecisionPointKo":"전용 클라이언트로 질문해 도구·턴 라우팅·이메일 분류 예제에 활용한다.","highlightBenefitJa":"型付き判断と生成モデルの Agent ループを分けて接続する。","highlightBenefitKo":"타입 판단과 생성 모델의 Agent 루프를 분리해 연결한다.","claimStatusJa":"固定バージョンの文書とソースを確認。実行・性能・安全性は当サイトで独立検証していません。","claimStatusKo":"고정 버전 문서와 소스를 확인했습니다. 실행·성능·안전성은 이 사이트에서 독립적으로 검증하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/bensyverson/goodall/blob/a612f04a712cf6c68926b780bd98e5087c687224/README.md","note":"README.md"},{"url":"https://github.com/bensyverson/goodall/blob/a612f04a712cf6c68926b780bd98e5087c687224/typesafe/client.go","note":"typesafe/client.go"}],"pinned":false},{"id":"gaurav-gosain:jev-go","name":"jev-go","author":"Gaurav-Gosain","url":"https://github.com/Gaurav-Gosain/jev-go","category":"SDK & Decision Frameworks","plainSummary":"Go 版 TypeSafe System One 客户端，提供类型化问题、答案与批量调用辅助。","plainSummaryEn":"A Go TypeSafe System One client with typed questions, answers and batching helpers.","jevDecisionPoint":"把应用状态和问题发送给 Jev，解析 Choice、Score 或 Noul 答案。","jevDecisionPointEn":"Sends application state and questions to Jev and parses Choice, Score or Noul answers.","highlightBenefit":"封装请求和响应解析，便于在应用中接入结构化判断。","highlightBenefitEn":"Packages requests and response parsing for typed decisions in applications.","tags":["multilanguage-sdk","typed-decisions"],"stars":3,"forks":0,"openIssues":0,"license":"MIT","language":"Go","lastCommitAt":"2026-09-16T05:55:49Z","createdAt":"2026-09-16T05:55:53Z","metadataFetchedAt":"2026-09-21T22:04:33.096Z","avatarUrl":"https://avatars.githubusercontent.com/u/55647468?v=4","summarySource":"human-reviewed","claimStatus":"已核对固定版本 README 与集成源码。","claimStatusEn":"README and integration source reviewed at a fixed commit; not independently run or benchmarked by this site.","plainSummaryJa":"型付き質問・回答と一括処理補助を備える Go の TypeSafe System One クライアント。","plainSummaryKo":"구조화된 질문·답변과 일괄 처리 도우미를 제공하는 Go TypeSafe System One 클라이언트다.","jevDecisionPointJa":"状態と質問を Jev に送り、Choice、Score、Noul の回答を解析する。","jevDecisionPointKo":"상태와 질문을 Jev에 보내 Choice, Score, Noul 답변을 처리한다.","highlightBenefitJa":"リクエストと応答解析をまとめ、アプリに型付き判断を組み込める。","highlightBenefitKo":"요청과 응답 처리를 묶어 앱에 구조화된 판단을 연결한다.","claimStatusJa":"固定版の README と実装を確認済み。このサイトでは独立した実行・性能検証をしていません。","claimStatusKo":"고정 버전 README와 구현을 확인했습니다. 이 사이트에서 독립적인 실행·성능 검증은 하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/Gaurav-Gosain/jev-go/blob/c9867e4afad0ddfaf4b9deb31038218495712fd0/client.go#L20","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/Gaurav-Gosain/jev-go/blob/c9867e4afad0ddfaf4b9deb31038218495712fd0/README.md","note":"固定版本的实现与说明 / Pinned source and documentation"}],"pinned":false},{"id":"ainame:swift-typesafe","name":"swift-typesafe","author":"ainame","url":"https://github.com/ainame/swift-typesafe","category":"SDK & Decision Frameworks","plainSummary":"社区 Swift TypeSafe 客户端，提供类型化问题、动态问题与响应解析。","plainSummaryEn":"A community Swift TypeSafe client with typed questions, dynamic questions and response parsing.","jevDecisionPoint":"通过 systemOne 发送 Jev 问题，用 Swift 类型或动态映射表示答案。","jevDecisionPointEn":"Sends Jev questions through systemOne and exposes answers as Swift types or dynamic maps.","highlightBenefit":"可在支持的平台上集成；具体 Swift 与系统版本要求见固定版本 README。","highlightBenefitEn":"Supports integration on its documented platforms; Swift and OS requirements are version-specific.","tags":["multilanguage-sdk","typed-decisions"],"stars":10,"forks":1,"openIssues":1,"license":"MIT","language":"Swift","lastCommitAt":"2026-09-20T08:02:53Z","createdAt":"2026-09-18T00:56:38Z","metadataFetchedAt":"2026-09-21T22:04:36.071Z","avatarUrl":"https://avatars.githubusercontent.com/u/748949?v=4","summarySource":"human-reviewed","claimStatus":"已核对固定版本 README 与调用源码。","claimStatusEn":"Reviewed the fixed-version README and call path.","plainSummaryJa":"型付き質問・動的質問・応答解析に対応するコミュニティ Swift TypeSafe クライアント。","plainSummaryKo":"타입 질문·동적 질문·응답 파싱을 지원하는 커뮤니티 Swift TypeSafe 클라이언트.","jevDecisionPointJa":"systemOne で Jev に問い、Swift 型または動的マップで回答を扱う。","jevDecisionPointKo":"systemOne으로 Jev에 질문하고 Swift 타입이나 동적 맵으로 응답을 다룬다.","highlightBenefitJa":"文書記載の環境に対応し、必要な Swift・OS 版は対象バージョンで確認する。","highlightBenefitKo":"문서에 명시된 플랫폼에 연동하며 Swift·OS 요구 버전은 해당 버전에서 확인한다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/ainame/swift-typesafe/blob/c990db869c10030b98af02e1219f5bca1e2c7cc2/README.md","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/ainame/swift-typesafe/blob/c990db869c10030b98af02e1219f5bca1e2c7cc2/Sources/TypeSafe/TypeSafeClient.swift","note":"固定版本说明与源码；本站未执行该项目。"}],"pinned":false},{"id":"nshkrdotcom:typesafe_sdk","name":"typesafe_sdk","author":"nshkrdotcom","url":"https://github.com/nshkrdotcom/typesafe_sdk","category":"SDK & Decision Frameworks","plainSummary":"面向 Elixir 的 TypeSafe SDK，把 Jev 的类型化问题与概率答案接到 Elixir 应用。","plainSummaryEn":"An Elixir TypeSafe SDK that brings Jev’s typed questions and probabilistic answers into Elixir applications.","jevDecisionPoint":"构造状态和 Noul、Choice、Score 请求，解析 System One 的回答。","jevDecisionPointEn":"Builds state and Noul, Choice, and Score requests and parses System One responses.","highlightBenefit":"复用 Elixir 数据结构与客户端封装接入判断接口。","highlightBenefitEn":"Reuses Elixir data structures and client wrappers for decision APIs.","tags":["multilanguage-sdk","typed-decisions"],"stars":5,"forks":1,"openIssues":0,"license":"MIT","language":"Elixir","lastCommitAt":"2026-09-20T03:39:38Z","createdAt":"2026-09-17T04:02:33Z","metadataFetchedAt":"2026-09-21T22:04:39.016Z","avatarUrl":"https://avatars.githubusercontent.com/u/127063941?v=4","summarySource":"human-reviewed","claimStatus":"当前 README 定位为 TypeSafe/Jev SDK；旧元数据里的通用 LLM 流式输出描述不宜直接当成已核实功能。","claimStatusEn":"The current README describes a TypeSafe/Jev SDK. Older metadata about generic LLM streaming should not be treated as verified functionality. This site has not run it.","plainSummaryJa":"Jev の型付き質問と確率的回答を Elixir アプリに組み込む TypeSafe SDK。","plainSummaryKo":"Jev의 타입 질문과 확률 답변을 Elixir 애플리케이션에 연결하는 TypeSafe SDK.","jevDecisionPointJa":"状態と Noul、Choice、Score のリクエストを作り、System One の応答を解析する。","jevDecisionPointKo":"상태와 Noul, Choice, Score 요청을 구성하고 System One 응답을 해석한다.","highlightBenefitJa":"Elixir のデータ構造とクライアントを再利用できる。","highlightBenefitKo":"Elixir 데이터 구조와 클라이언트 래퍼를 재사용한다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"confirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/nshkrdotcom/typesafe_sdk/blob/deaedd63882045ea0d8e9273002280a20809e170/lib/typesafe_sdk/generated/system_one.ex","note":"Fixed-commit source reviewed / 已核对固定版本源码"},{"url":"https://github.com/nshkrdotcom/typesafe_sdk/blob/deaedd63882045ea0d8e9273002280a20809e170/README.md","note":"Fixed-commit source reviewed / 已核对固定版本源码"}],"pinned":false,"topics":["agents","ai","ai-sdk","artificial-intelligence","elixir","function-calling","generative-ai","jev","llm","machine-learning","nshkr-ai-sdk","otp","sdk","streaming","structured-outputs","system-one","system-one-models","tool-calling","typesafe"]},{"id":"twister915:typesafe-ai","name":"typesafe-ai","author":"Twister915","url":"https://github.com/Twister915/typesafe-ai","category":"SDK & Decision Frameworks","plainSummary":"一个 Rust TypeSafe 客户端，提供异步 reqwest 或阻塞 ureq 后端，并可观察重试过程。","plainSummaryEn":"A Rust TypeSafe client with asynchronous reqwest or blocking ureq backends and observable retries.","jevDecisionPoint":"提交同一状态下的多个 Jev 问题，解析枚举形式的回答、概率和用量。","jevDecisionPointEn":"Sends multiple Jev questions over shared state and parses typed answers, probabilities and usage.","highlightBenefit":"可按应用选择同步或异步调用，同时保留可检查的错误信息。","highlightBenefitEn":"Offers synchronous or asynchronous integration with inspectable errors.","tags":["multilanguage-sdk","typed-decisions"],"stars":11,"forks":2,"openIssues":2,"license":"Apache-2.0","language":"Rust","lastCommitAt":"2026-09-16T13:23:45Z","createdAt":"2026-09-16T00:49:21Z","metadataFetchedAt":"2026-09-21T22:04:42.017Z","avatarUrl":"https://avatars.githubusercontent.com/u/2097155?v=4","summarySource":"human-reviewed","claimStatus":"已核对 README 与调用示例；未验证所有 feature 组合、后端或网络重试行为。","claimStatusEn":"The README and call example were inspected; all feature combinations, backends and retry behaviors were not independently tested.","plainSummaryJa":"非同期 reqwest または同期 ureq と、観測可能な再試行を備えた Rust TypeSafe クライアント。","plainSummaryKo":"비동기 reqwest 또는 동기 ureq와 확인 가능한 재시도를 제공하는 Rust TypeSafe 클라이언트.","jevDecisionPointJa":"共有状態に複数の Jev 質問を送り、回答・確率・用量を解析する。","jevDecisionPointKo":"공유 상태에 여러 Jev 질문을 보내 답·확률·사용량을 해석한다.","highlightBenefitJa":"同期・非同期を選択し、エラー情報を確認できる。","highlightBenefitKo":"동기·비동기 연동을 선택하고 오류 정보를 확인할 수 있다.","claimStatusJa":"固定バージョンの文書とソースを確認。実行・性能・安全性は当サイトで独立検証していません。","claimStatusKo":"고정 버전 문서와 소스를 확인했습니다. 실행·성능·안전성은 이 사이트에서 독립적으로 검증하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/Twister915/typesafe-ai/blob/d4455efb1d061ae6aac47b40c42ef390182201b1/README.md","note":"README.md"},{"url":"https://github.com/Twister915/typesafe-ai/blob/d4455efb1d061ae6aac47b40c42ef390182201b1/examples/tsg/main.rs","note":"examples/tsg/main.rs"},{"url":"https://github.com/Twister915/typesafe-ai/blob/d4455efb1d061ae6aac47b40c42ef390182201b1/src/reqwest_client.rs","note":"src/reqwest_client.rs"}],"pinned":false},{"id":"gilljon:typesafe-ai-rs","name":"typesafe-ai-rs","author":"gilljon","url":"https://github.com/gilljon/typesafe-ai-rs","category":"SDK & Decision Frameworks","plainSummary":"独立维护的 Rust SDK，提供异步与阻塞客户端、重试和响应元数据。","plainSummaryEn":"An independently maintained Rust SDK with async and blocking clients, retries and response metadata.","jevDecisionPoint":"把应用状态和问题发送给 Jev，解析 Choice、Score 或 Noul 答案。","jevDecisionPointEn":"Sends application state and questions to Jev and parses Choice, Score or Noul answers.","highlightBenefit":"封装请求和响应解析，便于在应用中接入结构化判断。","highlightBenefitEn":"Packages requests and response parsing for typed decisions in applications.","tags":["multilanguage-sdk","typed-decisions"],"stars":4,"forks":0,"openIssues":1,"license":"MIT","language":"Rust","lastCommitAt":"2026-09-17T05:13:49Z","createdAt":"2026-09-17T04:42:13Z","metadataFetchedAt":"2026-09-21T22:04:45.045Z","avatarUrl":"https://avatars.githubusercontent.com/u/113929785?v=4","summarySource":"human-reviewed","claimStatus":"已核对固定版本 README 与集成源码。","claimStatusEn":"README and integration source reviewed at a fixed commit; not independently run or benchmarked by this site.","plainSummaryJa":"非同期・同期クライアント、再試行、応答メタデータを備える独立開発の Rust SDK。","plainSummaryKo":"비동기·블로킹 클라이언트, 재시도와 응답 메타데이터를 제공하는 독립 Rust SDK다.","jevDecisionPointJa":"状態と質問を Jev に送り、Choice、Score、Noul の回答を解析する。","jevDecisionPointKo":"상태와 질문을 Jev에 보내 Choice, Score, Noul 답변을 처리한다.","highlightBenefitJa":"リクエストと応答解析をまとめ、アプリに型付き判断を組み込める。","highlightBenefitKo":"요청과 응답 처리를 묶어 앱에 구조화된 판단을 연결한다.","claimStatusJa":"固定版の README と実装を確認済み。このサイトでは独立した実行・性能検証をしていません。","claimStatusKo":"고정 버전 README와 구현을 확인했습니다. 이 사이트에서 독립적인 실행·성능 검증은 하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/gilljon/typesafe-ai-rs/blob/06f52208c22f63326226446926859174c538e296/src/client.rs#L78","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/gilljon/typesafe-ai-rs/blob/06f52208c22f63326226446926859174c538e296/README.md","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/gilljon/typesafe-ai-rs/blob/06f52208c22f63326226446926859174c538e296/src/lib.rs","note":"固定版本的实现与说明 / Pinned source and documentation"}],"pinned":false,"topics":["ai","api-client","rust","rust-sdk","typesafe"]},{"id":"joshmn:typesafe-sdk","name":"typesafe-sdk","author":"joshmn","url":"https://github.com/joshmn/typesafe-sdk","category":"SDK & Decision Frameworks","plainSummary":"TypeSafe System One 的社区 Ruby 客户端，默认使用 jev-latest。","plainSummaryEn":"A community Ruby client for TypeSafe System One, defaulting to jev-latest.","jevDecisionPoint":"用 Choice、Score、Noul 构造问题，发送请求后按答案类型提供访问接口。","jevDecisionPointEn":"Builds Choice, Score and Noul questions and exposes parsed answers by type.","highlightBenefit":"Ruby 程序可直接读取选择、分数和概率。","highlightBenefitEn":"Lets Ruby programs read choices, scores and probabilities directly.","tags":["multilanguage-sdk","typed-decisions"],"stars":6,"forks":3,"openIssues":6,"license":"MIT","language":"Ruby","lastCommitAt":"2026-09-19T19:38:30Z","createdAt":"2026-09-16T13:53:31Z","metadataFetchedAt":"2026-09-21T22:04:48.053Z","avatarUrl":"https://avatars.githubusercontent.com/u/2028176?v=4","summarySource":"human-reviewed","claimStatus":"已核对固定版本 README 与调用源码。","claimStatusEn":"Reviewed the fixed-version README and call path.","plainSummaryJa":"既定で jev-latest を使う TypeSafe System One のコミュニティ Ruby クライアント。","plainSummaryKo":"기본 모델로 jev-latest를 사용하는 TypeSafe System One 커뮤니티 Ruby 클라이언트.","jevDecisionPointJa":"Choice・Score・Noul で質問を作り、解析した回答を型別に取得する。","jevDecisionPointKo":"Choice·Score·Noul 질문을 만들고 파싱한 응답을 유형별로 제공한다.","highlightBenefitJa":"Ruby から選択・スコア・確率を直接読める。","highlightBenefitKo":"Ruby에서 선택지·점수·확률을 직접 읽을 수 있다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/joshmn/typesafe-sdk/blob/abde35d2ebae5d7dda900d1aa122ea5b653f1d9e/README.md","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/joshmn/typesafe-sdk/blob/abde35d2ebae5d7dda900d1aa122ea5b653f1d9e/lib/typesafe/sdk/client.rb","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/joshmn/typesafe-sdk/blob/abde35d2ebae5d7dda900d1aa122ea5b653f1d9e/lib/typesafe/sdk/net_http_transport.rb","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/joshmn/typesafe-sdk/blob/abde35d2ebae5d7dda900d1aa122ea5b653f1d9e/lib/typesafe/sdk/answer_parser.rb","note":"固定版本说明与源码；本站未执行该项目。"}],"pinned":false},{"id":"tangerg:typesafe-sdk-go","name":"typesafe-sdk-go","author":"Tangerg","url":"https://github.com/Tangerg/typesafe-sdk-go","category":"SDK & Decision Frameworks","plainSummary":"一个 Go TypeSafe SDK，用 Go 数据类型定义问题并读取 Jev 的选择、分数与概率。","plainSummaryEn":"A Go TypeSafe SDK for defining typed questions and reading Jev choices, scores, and probabilities.","jevDecisionPoint":"校验请求后调用 System One，将响应解码为类型化答案。","jevDecisionPointEn":"Validates requests, calls System One, and decodes typed answers.","highlightBenefit":"把鉴权、请求和错误处理封装进可复用的 Go 客户端。","highlightBenefitEn":"Packages authentication, requests, and error handling in a reusable Go client.","tags":["multilanguage-sdk","typed-decisions"],"stars":8,"forks":1,"openIssues":0,"license":"MIT","language":"Go","lastCommitAt":"2026-09-19T18:10:43Z","createdAt":"2026-09-18T09:11:50Z","metadataFetchedAt":"2026-09-21T22:04:51.044Z","avatarUrl":"https://avatars.githubusercontent.com/u/37281233?v=4","summarySource":"human-reviewed","claimStatus":"类型化返回值不保证概率校准或判断准确；未找到支持旧文案“专为分布式系统设计”的必要证据。","claimStatusEn":"Typed responses do not guarantee calibrated probabilities or correct judgments. The previous claim of being designed specifically for distributed systems was unsupported.","plainSummaryJa":"Go の型で質問を定義し、Jev の選択、点数、確率を読む TypeSafe SDK。","plainSummaryKo":"Go 타입으로 질문을 정의하고 Jev의 선택, 점수, 확률을 읽는 TypeSafe SDK.","jevDecisionPointJa":"リクエストを検証して System One を呼び、型付き回答に変換する。","jevDecisionPointKo":"요청을 검증하고 System One을 호출해 타입이 있는 답으로 해석한다.","highlightBenefitJa":"認証、リクエスト、エラー処理を Go クライアントで再利用できる。","highlightBenefitKo":"인증, 요청, 오류 처리를 Go 클라이언트에서 재사용한다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"confirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/Tangerg/typesafe-sdk-go/blob/e2f9f8353974fce1d2d2e4dd4283d5e306e8c2a8/client.go","note":"Fixed-commit source reviewed / 已核对固定版本源码"},{"url":"https://github.com/Tangerg/typesafe-sdk-go/blob/e2f9f8353974fce1d2d2e4dd4283d5e306e8c2a8/README.md","note":"Fixed-commit source reviewed / 已核对固定版本源码"}],"pinned":false},{"id":"keeltrace:hermes-jev","name":"hermes-jev","author":"keeltrace","url":"https://github.com/keeltrace/hermes-jev","category":"Decision Tools","plainSummary":"Hermes Agent 的异步 Jev 辅助决策系统，用于相关性、完成度、恢复路径和可选工具准入等判断。","plainSummaryEn":"An asynchronous Jev companion for Hermes Agent covering relevance, completion, recovery and optional admission decisions.","jevDecisionPoint":"Jev 在后台评估有边界的问题，Hermes 保留推理与执行职责，并记录判断来源。","jevDecisionPointEn":"Jev evaluates bounded questions in the background while Hermes retains reasoning and execution responsibilities.","highlightBenefit":"让模型判断有可追踪回执，同时可按配置减少对主流程的阻塞。","highlightBenefitEn":"Records decision provenance and supports configuration that avoids blocking ordinary execution.","tags":["mcp-integrations","evaluation-benchmarks","typed-decisions"],"stars":12,"forks":0,"openIssues":2,"license":"MIT","language":"Python","lastCommitAt":"2026-09-20T16:34:55Z","createdAt":"2026-09-18T02:10:22Z","metadataFetchedAt":"2026-09-21T22:04:54.004Z","avatarUrl":"https://avatars.githubusercontent.com/u/218190424?v=4","summarySource":"human-reviewed","claimStatus":"社区集成，非 TypeSafe 或 Nous 官方背书；不存在已核验的长任务准确率提升保证。","claimStatusEn":"A community integration without TypeSafe or Nous endorsement; improved long-task success rates are not established here. Not run here.","plainSummaryJa":"関連性・完了・復旧・任意の許可判断を補助する、Hermes Agent 用の非同期 Jev 連携。","plainSummaryKo":"관련성·완료·복구·선택형 허가 판단을 돕는 Hermes Agent용 비동기 Jev 연동.","jevDecisionPointJa":"Jev が限定した質問を背景で評価し、推論と実行は Hermes が担う。","jevDecisionPointKo":"Jev가 범위가 있는 질문을 백그라운드에서 평가하며 추론·실행은 Hermes가 맡는다.","highlightBenefitJa":"判断元を記録し、通常処理を妨げない設定を選べる。","highlightBenefitKo":"판단 출처를 기록하고 일반 처리를 막지 않는 설정을 선택한다.","claimStatusJa":"固定バージョンの文書とソースを確認。実行・性能・安全性は当サイトで独立検証していません。","claimStatusKo":"고정 버전 문서와 소스를 확인했습니다. 실행·성능·안전성은 이 사이트에서 독립적으로 검증하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/keeltrace/hermes-jev/blob/4feea5ef45aeb301622f18175ed4cf2e068b99bd/README.md","note":"README.md"},{"url":"https://github.com/keeltrace/hermes-jev/blob/4feea5ef45aeb301622f18175ed4cf2e068b99bd/hermes_jev/client.py","note":"hermes_jev/client.py"}],"pinned":false},{"id":"ying-kai-liao:jev-browser","name":"jev-browser","author":"Ying-Kai-Liao","url":"https://github.com/Ying-Kai-Liao/jev-browser","category":"Browser & OS Action","plainSummary":"浏览器自动化库、CLI 和 MCP：调用方模型给出目标与待输入文字，Jev 选择具体操作。","plainSummaryEn":"A browser library, CLI and MCP server where the caller supplies goals and text while Jev selects actions.","jevDecisionPoint":"从页面候选中选择元素、动作和值，并评估完成、错误与不可逆操作。","jevDecisionPointEn":"Chooses elements, actions and values, and assesses completion, errors and irreversible steps.","highlightBenefit":"把浏览器动作循环与调用方规划分开，并返回状态与轨迹。","highlightBenefitEn":"Separates browser-action loops from caller planning and returns status and traces.","tags":["browser-automation","mcp-integrations","cli-git-gates"],"stars":66,"forks":9,"openIssues":5,"license":"MIT","language":"JavaScript","lastCommitAt":"2026-09-18T13:23:08Z","createdAt":"2026-09-16T15:13:23Z","metadataFetchedAt":"2026-09-21T22:04:57.000Z","avatarUrl":"https://avatars.githubusercontent.com/u/62387690?v=4","summarySource":"human-reviewed","claimStatus":"40/42 任务结果和约 300ms 时延为作者指定站点测试；Token 对比是估算，不保证普遍成功率或费用收益。 已核对固定版本 README 与集成源码。","claimStatusEn":"The 40/42 result and about 300 ms per call are author-reported site-specific tests. Token comparisons are estimates, not universal success-rate or cost guarantees. README and integration source reviewed at a fixed commit; not independently run or benchmarked by this site.","plainSummaryJa":"呼び出し側が目標と入力文を渡し、Jev が操作を選ぶブラウザライブラリ、CLI、MCP サーバー。","plainSummaryKo":"호출 측이 목표와 입력 문장을 제공하고 Jev가 동작을 고르는 브라우저 라이브러리·CLI·MCP다.","jevDecisionPointJa":"要素、操作、値を選び、完了、エラー、取り消せない操作を評価する。","jevDecisionPointKo":"요소, 동작과 값을 고르고 완료, 오류와 되돌릴 수 없는 단계를 평가한다.","highlightBenefitJa":"操作ループを計画から分け、状態と履歴を返す。","highlightBenefitKo":"브라우저 동작 루프와 계획을 분리하며 상태와 기록을 반환한다.","claimStatusJa":"固定版の README と実装を確認済み。このサイトでは独立した実行・性能検証をしていません。","claimStatusKo":"고정 버전 README와 구현을 확인했습니다. 이 사이트에서 독립적인 실행·성능 검증은 하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/Ying-Kai-Liao/jev-browser/blob/578cff6e701a131733d03256078bb559a45ad188/src/jev.mjs#L8","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/Ying-Kai-Liao/jev-browser/blob/578cff6e701a131733d03256078bb559a45ad188/README.md","note":"固定版本的实现与说明 / Pinned source and documentation"}],"pinned":false,"topics":["browser-automation","llm-agents","mcp","mcp-server","playwright"]},{"id":"romaluev:jev-ego","name":"jev-ego","author":"romaluev","url":"https://github.com/romaluev/jev-ego","category":"Browser & OS Action","plainSummary":"面向 ego lite 的浏览器 Agent，把页面可操作元素编号后交给 Jev 选择下一步。","plainSummaryEn":"A browser Agent for ego lite that numbers actionable elements for Jev to choose the next step.","jevDecisionPoint":"Jev 在一次请求中选择操作及目标；需要自由文本时使用单独的文字辅助模型。","jevDecisionPointEn":"Jev selects an operation and target in one request; a separate helper handles free-form text when needed.","highlightBenefit":"提供观察、建议和执行入口；上传、弹窗等能力仍需其他浏览器工具。","highlightBenefitEn":"Offers observation, suggestions and execution, while uploads and dialogs still require other browser tools.","tags":["browser-automation","typed-decisions"],"stars":13,"forks":0,"openIssues":0,"license":null,"language":"TypeScript","lastCommitAt":"2026-09-17T14:35:26Z","createdAt":"2026-09-17T14:35:09Z","metadataFetchedAt":"2026-09-21T22:05:00.046Z","avatarUrl":"https://avatars.githubusercontent.com/u/109923636?v=4","summarySource":"human-reviewed","claimStatus":"已核对固定版本 README 与调用源码。","claimStatusEn":"Reviewed the fixed-version README and call path.","plainSummaryJa":"ego lite の操作要素を番号付き一覧にし、Jev が次の動作を選ぶブラウザー Agent。","plainSummaryKo":"ego lite의 조작 가능한 요소에 번호를 붙여 Jev가 다음 동작을 고르는 브라우저 Agent.","jevDecisionPointJa":"一度の要求で操作と対象を選び、自由文が必要なら別の補助モデルを使う。","jevDecisionPointKo":"한 번의 요청으로 동작·대상을 고르고 자유 문장은 별도 보조 모델이 처리한다.","highlightBenefitJa":"観測・提案・実行に対応するが、アップロードやダイアログは別のツールを使う。","highlightBenefitKo":"관측·제안·실행을 제공하지만 업로드·대화상자는 다른 브라우저 도구가 필요하다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/romaluev/jev-ego/blob/12eaafe795db56da821d9268ee7b9ae9ea24b23b/README.md","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/romaluev/jev-ego/blob/12eaafe795db56da821d9268ee7b9ae9ea24b23b/src/model.ts","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/romaluev/jev-ego/blob/12eaafe795db56da821d9268ee7b9ae9ea24b23b/src/agent.ts","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/romaluev/jev-ego/blob/12eaafe795db56da821d9268ee7b9ae9ea24b23b/LICENSE","note":"固定版本说明与源码；本站未执行该项目。"}],"pinned":false,"topics":["ai-agent","browser-agent","cdp","cli","ego-browser","ego-lite","jev","typesafe","typesafe-ai","typescript","web-automation"]},{"id":"giuliosmall:pg_typesafe","name":"pg_typesafe","author":"giuliosmall","url":"https://github.com/giuliosmall/pg_typesafe","category":"Data & Search","plainSummary":"一个预览阶段的 PostgreSQL C 扩展，让 SQL 直接调用 Jev 做分类、是非判断和评分。","plainSummaryEn":"A pre-alpha PostgreSQL C extension for calling Jev from SQL for classification, yes/no judgments, and scoring.","jevDecisionPoint":"将 SQL 输入转换为 System One 请求，并把答案返回给数据库函数。","jevDecisionPointEn":"Converts SQL input into System One requests and returns answers through database functions.","highlightBenefit":"在现有数据库查询中加入类型化语义判断。","highlightBenefitEn":"Adds typed semantic judgments to existing database queries.","tags":["database-vector","classification-ranking"],"stars":81,"forks":1,"openIssues":0,"license":"MIT","language":"C","lastCommitAt":"2026-09-20T10:52:41Z","createdAt":"2026-09-17T14:45:16Z","metadataFetchedAt":"2026-09-21T22:05:03.031Z","avatarUrl":"https://avatars.githubusercontent.com/u/49522377?v=4","summarySource":"human-reviewed","claimStatus":"作者标注 pre-alpha；需要安装数据库扩展并配置权限，数据发送到外部 API。本站未做数据库运行验证。","claimStatusEn":"Marked pre-alpha by the author. Requires extension installation and permissions; data is sent to an external API. This site has not tested it in a database.","plainSummaryJa":"SQL から Jev の分類、二択、採点を呼び出す pre-alpha の PostgreSQL C 拡張。","plainSummaryKo":"SQL에서 Jev의 분류, 예·아니요 판단, 평가를 호출하는 pre-alpha PostgreSQL C 확장.","jevDecisionPointJa":"SQL 入力を System One リクエストに変え、データベース関数で回答を返す。","jevDecisionPointKo":"SQL 입력을 System One 요청으로 바꾸고 데이터베이스 함수로 답을 반환한다.","highlightBenefitJa":"既存の問い合わせに型付きの意味判断を加える。","highlightBenefitKo":"기존 쿼리에 타입이 있는 의미 판단을 추가한다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"confirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/giuliosmall/pg_typesafe/blob/4b5bfc1df11b18c3f07bb10804eeb47e4508ec6a/typesafe.c","note":"Fixed-commit source reviewed / 已核对固定版本源码"},{"url":"https://github.com/giuliosmall/pg_typesafe/blob/4b5bfc1df11b18c3f07bb10804eeb47e4508ec6a/README.md","note":"Fixed-commit source reviewed / 已核对固定版本源码"}],"pinned":false},{"id":"kieranklaassen:ruby_llm-typesafe","name":"ruby_llm-typesafe","author":"kieranklaassen","url":"https://github.com/kieranklaassen/ruby_llm-typesafe","category":"SDK & Integrations","plainSummary":"为 RubyLLM 2 添加 TypeSafe provider，通过结构化输出接口调用 Jev 的三类判断。","plainSummaryEn":"A TypeSafe provider for RubyLLM 2 that exposes Jev’s three judgment types through structured output.","jevDecisionPoint":"用 Schema 构造 Noul、Choice、Score 问题，再把结果交回 RubyLLM 应用。","jevDecisionPointEn":"Builds Noul, Choice and Score questions with Schema and returns answers to the RubyLLM application.","highlightBenefit":"在现有 RubyLLM 程序里复用明确的结构化判断接口。","highlightBenefitEn":"Reuses the structured-output interface in existing RubyLLM programs.","tags":["multilanguage-sdk","typed-decisions"],"stars":18,"forks":0,"openIssues":2,"license":"MIT","language":"Ruby","lastCommitAt":"2026-09-16T05:05:32Z","createdAt":"2026-09-16T04:03:33Z","metadataFetchedAt":"2026-09-21T22:05:06.076Z","avatarUrl":"https://avatars.githubusercontent.com/u/209089?v=4","summarySource":"human-reviewed","claimStatus":"只支持该 provider 声明的结构化输出路径，不是通用聊天功能替换。","claimStatusEn":"Limited to the provider’s supported structured-output path, not a general chat replacement.","plainSummaryJa":"構造化出力を通じて Jev の三種の判断を扱う、RubyLLM 2 用 TypeSafe provider。","plainSummaryKo":"구조화 출력을 통해 Jev의 세 판단 유형을 제공하는 RubyLLM 2용 TypeSafe provider.","jevDecisionPointJa":"Schema で Noul・Choice・Score を作り、回答を RubyLLM アプリへ返す。","jevDecisionPointKo":"Schema로 Noul·Choice·Score를 만들고 답을 RubyLLM 앱에 반환한다.","highlightBenefitJa":"既存 RubyLLM プログラムで構造化判断のインターフェースを使える。","highlightBenefitKo":"기존 RubyLLM 프로그램에서 구조화 판단 인터페이스를 재사용한다.","claimStatusJa":"固定バージョンの文書とソースを確認。実行・性能・安全性は当サイトで独立検証していません。","claimStatusKo":"고정 버전 문서와 소스를 확인했습니다. 실행·성능·안전성은 이 사이트에서 독립적으로 검증하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/kieranklaassen/ruby_llm-typesafe/blob/33e680115a43d584b1c0dcff385816f459a4bc71/README.md","note":"README.md"},{"url":"https://github.com/kieranklaassen/ruby_llm-typesafe/blob/33e680115a43d584b1c0dcff385816f459a4bc71/lib/ruby_llm/providers/typesafe.rb","note":"lib/ruby_llm/providers/typesafe.rb"}],"pinned":false},{"id":"akashpriyadarshii:jev-seo","name":"jev-seo","author":"AkashPriyadarshii","url":"https://github.com/AkashPriyadarshii/jev-seo","category":"Domain & Vertical Tools","plainSummary":"Rust SEO/GEO 命令行与 MCP 实验工具，结合网页检查、DuckDuckGo 查询和可选 Jev 评分。","plainSummaryEn":"An experimental Rust SEO/GEO CLI and MCP tool combining page checks, DuckDuckGo queries and optional Jev scoring.","jevDecisionPoint":"给搜索意图、内容直接性和内容缺口分类，并按自定义 rubric 估计可引用性。","jevDecisionPointEn":"Classifies search intent, direct answers and content gaps, with a custom citation-likelihood rubric.","highlightBenefit":"把本地检查、搜索结果和模型评分整理成可查看的报告。","highlightBenefitEn":"Collects local checks, search results and model judgments into reports.","tags":["domain-workflows","cli-git-gates","mcp-integrations"],"stars":29,"forks":3,"openIssues":1,"license":"MIT","language":"Rust","lastCommitAt":"2026-09-21T17:52:55Z","createdAt":"2026-09-18T13:02:32Z","metadataFetchedAt":"2026-09-21T22:05:09.001Z","avatarUrl":"https://avatars.githubusercontent.com/u/272530059?v=4","summarySource":"human-reviewed","claimStatus":"Jev 需要 API Key，不能宣称完全免费；GEO 分数是模型 rubric，不是实际搜索引擎收录、引用或排名测量。 已核对固定版本 README 与集成源码。","claimStatusEn":"Jev requires an API key, so this is not unconditionally free. The GEO score is a model rubric, not measured search-engine indexing, citation or ranking. README and integration source reviewed at a fixed commit; not independently run or benchmarked by this site.","plainSummaryJa":"ページ検査、DuckDuckGo 検索、任意の Jev 評価を組み合わせる実験的な Rust SEO/GEO CLI と MCP。","plainSummaryKo":"페이지 검사, DuckDuckGo 검색과 선택형 Jev 평가를 결합한 실험적 Rust SEO/GEO CLI 및 MCP다.","jevDecisionPointJa":"検索意図、直接的回答、内容不足を分類し、独自尺度で引用されやすさを推定する。","jevDecisionPointKo":"검색 의도, 직접 답변과 내용 부족을 분류하고 자체 기준으로 인용 가능성을 추정한다.","highlightBenefitJa":"ローカル検査、検索結果、モデル判断をレポートにまとめる。","highlightBenefitKo":"로컬 검사, 검색 결과와 모델 판단을 보고서로 정리한다.","claimStatusJa":"固定版の README と実装を確認済み。このサイトでは独立した実行・性能検証をしていません。","claimStatusKo":"고정 버전 README와 구현을 확인했습니다. 이 사이트에서 독립적인 실행·성능 검증은 하지 않았습니다.","licenseStatus":"unconfirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/AkashPriyadarshii/jev-seo/blob/bbd59fe994d557e12f31d3548a012e0f02db64d9/src/engine.rs#L28","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/AkashPriyadarshii/jev-seo/blob/bbd59fe994d557e12f31d3548a012e0f02db64d9/README.md","note":"固定版本的实现与说明 / Pinned source and documentation"}],"pinned":false,"topics":["ahrefs-alternative","claude-code","cli","duckduckgo","generative-engine-optimization","geo","jev","mcp","rank-tracker","rust","search-engine-optimization","semrush-alternative","seo","serp","typesafe","typesafe-ai"]},{"id":"baronunread:leanest","name":"leanest","author":"baronunread","url":"https://github.com/baronunread/leanest","category":"Codebase & Graph Pathfinding","plainSummary":"在现有测试运行器前增加 Jev 筛选，根据 diff 和测试源码判断哪些测试应运行。","plainSummaryEn":"Adds Jev-based selection before an existing test runner using diffs and test source.","jevDecisionPoint":"Jev 评估测试与变更的关系，本地策略在不确定、调用失败或测试文件改动时选择运行。","jevDecisionPointEn":"Jev judges relevance; local policy runs tests on uncertainty, API failure or changes to the test file itself.","highlightBenefit":"可用 shadow 模式对比建议；选择结果不能保证没有漏测。","highlightBenefitEn":"Supports shadow-mode comparison, but selected subsets cannot guarantee complete regression coverage.","tags":["coding-agents","ast-code-graph","evaluation-benchmarks"],"stars":3,"forks":0,"openIssues":0,"license":"MIT","language":"TypeScript","lastCommitAt":"2026-09-21T08:44:09Z","createdAt":"2026-09-17T14:25:25Z","metadataFetchedAt":"2026-09-21T22:05:12.011Z","avatarUrl":"https://avatars.githubusercontent.com/u/50637199?v=4","summarySource":"human-reviewed","claimStatus":"源码显示不确定时倾向运行；未测量漏测率或证明测试子集覆盖完整。","claimStatusEn":"Source shows conservative run behavior on uncertainty; missed-test rates and complete subset coverage were not measured.","plainSummaryJa":"diff とテストソースを使い、既存テストランナーの前に Jev 選別を加える。","plainSummaryKo":"diff와 테스트 소스로 기존 테스트 실행기 앞에 Jev 선별을 추가한다.","jevDecisionPointJa":"関連性を Jev が判断し、不確実・API 障害・テスト自身の変更時はローカル方針で実行する。","jevDecisionPointKo":"Jev가 관련성을 판단하고 불확실·API 실패·테스트 파일 변경 시 로컬 정책으로 실행한다.","highlightBenefitJa":"shadow モードで比較できるが、選別後のテストだけで漏れがないとは保証しない。","highlightBenefitKo":"shadow 모드로 비교할 수 있지만 선택된 테스트만으로 누락이 없다고 보장하지 않는다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/baronunread/leanest/blob/dfb8c6b83e5b362d8258ee64835e15b5766e08d2/README.md","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/baronunread/leanest/blob/dfb8c6b83e5b362d8258ee64835e15b5766e08d2/src/jev-client.ts","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/baronunread/leanest/blob/dfb8c6b83e5b362d8258ee64835e15b5766e08d2/src/sieve.ts","note":"固定版本说明与源码；本站未执行该项目。"}],"pinned":false},{"id":"iammrduncan:typesafe-ai-benchmark","name":"typesafe-ai-benchmark","author":"iammrduncan","url":"https://github.com/iammrduncan/typesafe-ai-benchmark","category":"Evaluation & Observability","plainSummary":"在共同的应用任务上对照 Jev 与其他结构化输出模型，记录错误、延迟、Token 与估算成本。","plainSummaryEn":"Compares Jev with other structured-output models on shared application tasks, recording errors, latency, Tokens, and estimated cost.","jevDecisionPoint":"把同一任务转换成 Jev 的 Choice/Noul 问题，并把答案映射到统一的结果格式。","jevDecisionPointEn":"Maps the same tasks to Jev Choice/Noul questions and normalizes answers to a shared result format.","highlightBenefit":"保留对照方法与结果，方便检查模型差异。","highlightBenefitEn":"Preserves comparison methods and results for inspecting model differences.","tags":["evaluation-benchmarks","typed-decisions"],"stars":37,"forks":5,"openIssues":1,"license":"MIT","language":"TypeScript","lastCommitAt":"2026-09-19T06:00:51Z","createdAt":"2026-09-16T21:45:09Z","metadataFetchedAt":"2026-09-21T22:05:15.017Z","avatarUrl":"https://avatars.githubusercontent.com/u/289292190?v=4","summarySource":"human-reviewed","claimStatus":"README 当前描述有限的合成工作负载与独立捕获结果；演示视频并非测量数据，不能推广成所有任务上的排名。本站未复现。","claimStatusEn":"The current README describes limited synthetic workloads and separately captured results. The demo video is not measurement data, and results are not a ranking for all tasks. This site has not reproduced them.","plainSummaryJa":"共通のタスクで Jev とほかの構造化出力モデルを比べ、誤り、遅延、Token、推定費用を記録する。","plainSummaryKo":"공통 작업에서 Jev와 다른 구조화 출력 모델을 비교하며 오류, 지연, Token, 추정 비용을 기록한다.","jevDecisionPointJa":"同じタスクを Choice/Noul に変換し、回答を共通の結果形式にそろえる。","jevDecisionPointKo":"같은 작업을 Choice/Noul로 변환하고 답을 공통 결과 형식으로 맞춘다.","highlightBenefitJa":"比較手順と結果からモデル間の違いを確認できる。","highlightBenefitKo":"비교 방법과 결과로 모델 간 차이를 살펴볼 수 있다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"confirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/iammrduncan/typesafe-ai-benchmark/blob/cf348cd291bb538ad9fa902334649ef5dd937e92/packages/demos/lib/jev.ts","note":"Fixed-commit source reviewed / 已核对固定版本源码"},{"url":"https://github.com/iammrduncan/typesafe-ai-benchmark/blob/cf348cd291bb538ad9fa902334649ef5dd937e92/README.md","note":"Fixed-commit source reviewed / 已核对固定版本源码"}],"pinned":false},{"id":"inanna-malick:jev-dsl","name":"jev-dsl","author":"inanna-malick","url":"https://github.com/inanna-malick/jev-dsl","category":"SDK & Decision Frameworks","plainSummary":"一个早期 Haskell DSL，用表达式描述带标签的 Jev 问题，生成请求并解析对应答案。","plainSummaryEn":"An early Haskell DSL that describes labeled Jev questions, renders requests and decodes matching answers.","jevDecisionPoint":"用类型推导与标签处理器把 Choice 答案接到事先定义的程序分支。","jevDecisionPointEn":"Uses inferred types and labeled handlers to connect Choice answers to predefined program branches.","highlightBenefit":"将问题、返回类型与分支代码放在同一套可检查定义中。","highlightBenefitEn":"Keeps questions, result types and branch handling in one inspectable definition.","tags":["multilanguage-sdk","typed-decisions"],"stars":7,"forks":0,"openIssues":0,"license":"MIT","language":"Haskell","lastCommitAt":"2026-09-18T03:29:55Z","createdAt":"2026-09-17T17:14:51Z","metadataFetchedAt":"2026-09-21T22:05:17.998Z","avatarUrl":"https://avatars.githubusercontent.com/u/2675660?v=4","summarySource":"human-reviewed","claimStatus":"README 标明 early alpha 且接口仍在变化；类型保证不等于判断事实正确，库本身不负责联网。","claimStatusEn":"The README marks it early alpha with a changing interface. Type guarantees do not establish factual correctness, and the library itself has no network transport.","plainSummaryJa":"ラベル付き Jev 質問を記述し、要求生成と回答解析を行う初期 Haskell DSL。","plainSummaryKo":"라벨이 있는 Jev 질문을 기술하고 요청 생성·응답 해석을 하는 초기 Haskell DSL.","jevDecisionPointJa":"型推論とラベル付き処理器で、Choice の答えを定義済み分岐へ接続する。","jevDecisionPointKo":"타입 추론과 라벨 처리기로 Choice 답을 미리 정의한 분기에 연결한다.","highlightBenefitJa":"質問・戻り値の型・分岐処理を一つの定義で確認できる。","highlightBenefitKo":"질문·반환 타입·분기 처리를 하나의 정의에서 확인한다.","claimStatusJa":"固定バージョンの文書とソースを確認。実行・性能・安全性は当サイトで独立検証していません。","claimStatusKo":"고정 버전 문서와 소스를 확인했습니다. 실행·성능·안전성은 이 사이트에서 독립적으로 검증하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/inanna-malick/jev-dsl/blob/f16f1363b4d389d6e34f9d695fbd254ca0735f2e/README.md","note":"README.md"},{"url":"https://github.com/inanna-malick/jev-dsl/blob/f16f1363b4d389d6e34f9d695fbd254ca0735f2e/scripts/example.sh","note":"scripts/example.sh"}],"pinned":false},{"id":"zhirschtritt:typesafe-go","name":"typesafe-go","author":"zhirschtritt","url":"https://github.com/zhirschtritt/typesafe-go","category":"SDK & Decision Frameworks","plainSummary":"无第三方依赖的非官方 Go 客户端，支持 System One 请求和模型列表。","plainSummaryEn":"An unofficial Go client without third-party dependencies for System One calls and model discovery.","jevDecisionPoint":"把应用状态和问题发送给 Jev，解析 Choice、Score 或 Noul 答案。","jevDecisionPointEn":"Sends application state and questions to Jev and parses Choice, Score or Noul answers.","highlightBenefit":"封装请求和响应解析，便于在应用中接入结构化判断。","highlightBenefitEn":"Packages requests and response parsing for typed decisions in applications.","tags":["multilanguage-sdk","typed-decisions"],"stars":1,"forks":0,"openIssues":0,"license":"MIT","language":"Go","lastCommitAt":"2026-09-17T16:13:33Z","createdAt":"2026-09-16T21:47:51Z","metadataFetchedAt":"2026-09-21T22:05:21.020Z","avatarUrl":"https://avatars.githubusercontent.com/u/8853023?v=4","summarySource":"human-reviewed","claimStatus":"已核对固定版本 README 与集成源码。","claimStatusEn":"README and integration source reviewed at a fixed commit; not independently run or benchmarked by this site.","plainSummaryJa":"System One 呼び出しとモデル一覧に対応する、外部依存のない非公式 Go クライアント。","plainSummaryKo":"System One 호출과 모델 조회를 제공하는 외부 의존성 없는 비공식 Go 클라이언트다.","jevDecisionPointJa":"状態と質問を Jev に送り、Choice、Score、Noul の回答を解析する。","jevDecisionPointKo":"상태와 질문을 Jev에 보내 Choice, Score, Noul 답변을 처리한다.","highlightBenefitJa":"リクエストと応答解析をまとめ、アプリに型付き判断を組み込める。","highlightBenefitKo":"요청과 응답 처리를 묶어 앱에 구조화된 판단을 연결한다.","claimStatusJa":"固定版の README と実装を確認済み。このサイトでは独立した実行・性能検証をしていません。","claimStatusKo":"고정 버전 README와 구현을 확인했습니다. 이 사이트에서 독립적인 실행·성능 검증은 하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/zhirschtritt/typesafe-go/blob/a4062e525d1ad23aa739d3454247bcec52486a57/client.go#L18","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/zhirschtritt/typesafe-go/blob/a4062e525d1ad23aa739d3454247bcec52486a57/README.md","note":"固定版本的实现与说明 / Pinned source and documentation"}],"pinned":false,"topics":["ai","jev","sdk","systemone","typesafe"]},{"id":"qkal:canny","name":"Canny","author":"qkal","url":"https://github.com/qkal/Canny","category":"Evaluation & Observability","plainSummary":"为 Claude Code 与 Codex CLI 记录执行账本，检查改动后是否有通过的验证。","plainSummaryEn":"Keeps an execution ledger for Claude Code and Codex CLI to check for passing validation after edits.","jevDecisionPoint":"Jev 可识别完成声明和语义规则问题；阻止结束依赖账本事实与本地规则。","jevDecisionPointEn":"Jev can identify completion claims and semantic-rule issues; stop gates depend on ledger facts and local rules.","highlightBenefit":"区分执行证据与模型意见，Jev 不单独决定任务已通过验收。","highlightBenefitEn":"Separates execution evidence from model opinion rather than letting Jev alone certify completion.","tags":["coding-agents","cli-git-gates","evaluation-benchmarks"],"stars":27,"forks":2,"openIssues":0,"license":"MIT","language":"TypeScript","lastCommitAt":"2026-09-21T17:21:39Z","createdAt":"2026-09-11T20:48:57Z","metadataFetchedAt":"2026-09-21T22:05:24.001Z","avatarUrl":"https://avatars.githubusercontent.com/u/77361240?v=4","summarySource":"human-reviewed","claimStatus":"Jev 只提供辅助判断，阻断依据是记录的事实和本地规则；未验证能覆盖所有虚假完成声明。","claimStatusEn":"Jev supplies advisory judgments while recorded facts and local rules govern blocking; complete coverage of false completion claims is unverified.","plainSummaryJa":"Claude Code / Codex CLI の実行台帳から、変更後に検証が通ったか確認する。","plainSummaryKo":"Claude Code·Codex CLI 실행 장부로 수정 후 검증이 통과했는지 확인한다.","jevDecisionPointJa":"Jev は完了声明や意味ルールの問題を補助判断し、停止ゲートは台帳とローカル規則で決める。","jevDecisionPointKo":"Jev는 완료 주장·의미 규칙 문제를 보조 판단하고 종료 차단은 장부 사실과 로컬 규칙으로 결정한다.","highlightBenefitJa":"実行証拠とモデル意見を分け、Jev 単独で完了認定しない。","highlightBenefitKo":"실행 증거와 모델 의견을 구분하며 Jev 단독으로 완료를 인증하지 않는다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/qkal/Canny/blob/74bc3487370ae6d61cef69c5642cce504a8579cb/README.md","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/qkal/Canny/blob/74bc3487370ae6d61cef69c5642cce504a8579cb/src/jev.ts","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/qkal/Canny/blob/74bc3487370ae6d61cef69c5642cce504a8579cb/src/hook.ts","note":"固定版本说明与源码；本站未执行该项目。"}],"pinned":false,"topics":["agent-hooks","ai-agents","claude-code","claude-code-hooks","codex","codex-cli","guardrails","jev","typesafe","typescript"]},{"id":"devanshbatham:commit-miner","name":"commit-miner","author":"devanshbatham","url":"https://github.com/devanshbatham/commit-miner","category":"Codebase & Graph Pathfinding","plainSummary":"用 Jev 给 Git commit 的日志和 diff 分类，整理 bug 修复、安全修复、CWE 与改动类型。","plainSummaryEn":"Classifies Git commit messages and diffs with Jev for bug fixes, security fixes, CWEs, and change types.","jevDecisionPoint":"对提交内容询问固定类别问题，保存结果供过滤和 HTML/CSV 报告使用。","jevDecisionPointEn":"Asks fixed-category questions about commits and saves results for filtering and HTML/CSV reports.","highlightBenefit":"把大量提交整理成便于进一步核查的分类记录。","highlightBenefitEn":"Organizes large commit histories into records for further review.","tags":["ast-code-graph","cli-git-gates","classification-ranking"],"stars":33,"forks":6,"openIssues":0,"license":null,"language":"Rust","lastCommitAt":"2026-09-17T17:06:32Z","createdAt":"2026-09-17T08:00:01Z","metadataFetchedAt":"2026-09-21T22:05:27.001Z","avatarUrl":"https://avatars.githubusercontent.com/u/28016517?v=4","summarySource":"human-reviewed","claimStatus":"分类与 CWE 标签是模型判断，不能当作已确认漏洞；本站未验证准确率，提交内容会发送到外部 API。","claimStatusEn":"Classifications and CWE labels are model judgments, not confirmed vulnerabilities. This site has not verified accuracy; commit content is sent to an external API.","plainSummaryJa":"Git のメッセージと diff を Jev で分類し、バグ修正、安全性修正、CWE、変更種別を整理する。","plainSummaryKo":"Git 메시지와 diff를 Jev로 분류해 버그 수정, 보안 수정, CWE, 변경 유형을 정리한다.","jevDecisionPointJa":"固定カテゴリを質問し、フィルターや HTML/CSV レポート用に保存する。","jevDecisionPointKo":"고정 범주를 질문하고 필터나 HTML/CSV 보고서용으로 저장한다.","highlightBenefitJa":"大量の履歴を追加確認しやすい分類記録にする。","highlightBenefitKo":"많은 커밋 이력을 추가 검토하기 쉬운 분류 기록으로 만든다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"unconfirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/devanshbatham/commit-miner/blob/977617ebce07c56b965253a68577b1d92b93fdf1/src/jev.rs","note":"Fixed-commit source reviewed / 已核对固定版本源码"},{"url":"https://github.com/devanshbatham/commit-miner/blob/977617ebce07c56b965253a68577b1d92b93fdf1/README.md","note":"Fixed-commit source reviewed / 已核对固定版本源码"}],"pinned":false},{"id":"ilkerulusoy:pi-jev-compact","name":"pi-jev-compact","author":"ilkerulusoy","url":"https://github.com/ilkerulusoy/pi-jev-compact","category":"Context GC & Filter","plainSummary":"Pi 上下文整理扩展，默认筛除旧工具历史，另可选择把助手文字纳入候选。","plainSummaryEn":"A Pi context-pruning extension targeting tool history by default, with optional assistant-prose pruning.","jevDecisionPoint":"Jev 判断候选是否还需要保留，剩余文本按原样交回。","jevDecisionPointEn":"Jev judges which candidates remain useful; surviving text is returned verbatim.","highlightBenefit":"让删减范围可配置，并保留判断记录供检查。","highlightBenefitEn":"Provides configurable pruning scope and decision records.","tags":["context-compaction","coding-agents"],"stars":0,"forks":0,"openIssues":0,"license":null,"language":"TypeScript","lastCommitAt":"2026-09-18T15:10:26Z","createdAt":"2026-09-18T10:58:27Z","metadataFetchedAt":"2026-09-21T22:05:29.997Z","avatarUrl":"https://avatars.githubusercontent.com/u/1990736?v=4","summarySource":"human-reviewed","claimStatus":"作者明确表示未按参考答案测量保留/删除质量；合成窗口回收比例不是普遍效果，更不能保证不丢上下文。 已核对固定版本 README 与集成源码。","claimStatusEn":"The author has not evaluated keep/drop quality against reference judgments. Synthetic context reductions do not guarantee useful-context preservation. README and integration source reviewed at a fixed commit; not independently run or benchmarked by this site.","plainSummaryJa":"既定では古いツール履歴を、任意で助手の文章も整理する Pi 拡張。","plainSummaryKo":"기본적으로 도구 기록을 정리하고 선택적으로 어시스턴트 문장도 정리하는 Pi 확장이다.","jevDecisionPointJa":"候補を残す必要性を Jev が判断し、残る文章は原文のまま返す。","jevDecisionPointKo":"Jev가 후보의 필요성을 판단하고 남는 텍스트는 원문 그대로 반환한다.","highlightBenefitJa":"削減範囲を設定でき、判断の記録を確認できる。","highlightBenefitKo":"정리 범위를 설정하고 판단 기록을 확인할 수 있다.","claimStatusJa":"固定版の README と実装を確認済み。このサイトでは独立した実行・性能検証をしていません。","claimStatusKo":"고정 버전 README와 구현을 확인했습니다. 이 사이트에서 독립적인 실행·성능 검증은 하지 않았습니다.","licenseStatus":"unconfirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/ilkerulusoy/pi-jev-compact/blob/7d54b2e9a437fcf445c7f50aa49d673dbd863176/src/core/jev.ts#L3","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/ilkerulusoy/pi-jev-compact/blob/7d54b2e9a437fcf445c7f50aa49d673dbd863176/README.md","note":"固定版本的实现与说明 / Pinned source and documentation"}],"pinned":false},{"id":"kevinpita:pi-jev-context","name":"pi-jev-context","author":"kevinpita","url":"https://github.com/kevinpita/pi-jev-context","category":"Context GC & Filter","plainSummary":"Pi 的可逆上下文筛选扩展，用 Jev 判断旧消息是否仍值得发给主模型。","plainSummaryEn":"A reversible Pi context filter using Jev to judge whether older messages remain useful.","jevDecisionPoint":"对历史片段做相关性判断，低分片段从后续请求隐藏，原始会话保留。","jevDecisionPointEn":"Scores historical fragments and hides low-scoring content from future requests while preserving session history.","highlightBenefit":"可关闭筛选恢复完整上下文；启用时部分历史内容会发送给 TypeSafe。","highlightBenefitEn":"Disabling restores full context; enabling sends selected history to TypeSafe.","tags":["context-compaction","coding-agents"],"stars":2,"forks":0,"openIssues":0,"license":"MIT","language":"TypeScript","lastCommitAt":"2026-09-18T16:10:40Z","createdAt":"2026-09-18T16:04:30Z","metadataFetchedAt":"2026-09-21T22:05:33.042Z","avatarUrl":"https://avatars.githubusercontent.com/u/69063372?v=4","summarySource":"human-reviewed","claimStatus":"已核对固定版本 README 与调用源码。","claimStatusEn":"Reviewed the fixed-version README and call path.","plainSummaryJa":"古いメッセージの有用性を Jev で判断する Pi の可逆コンテキストフィルター。","plainSummaryKo":"오래된 메시지의 유용성을 Jev로 판단하는 Pi의 가역적 컨텍스트 필터.","jevDecisionPointJa":"履歴断片を評価し、低評価部分を今後の要求から隠すが会話原本は保持する。","jevDecisionPointKo":"기록 조각을 평가해 낮은 점수의 내용을 이후 요청에서 숨기지만 원본 세션은 보존한다.","highlightBenefitJa":"無効化で全コンテキストへ戻せる。有効時は一部履歴を TypeSafe へ送る。","highlightBenefitKo":"끄면 전체 컨텍스트를 복원하며 켜면 일부 기록을 TypeSafe로 전송한다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/kevinpita/pi-jev-context/blob/640ffa74c376fe8e487e4fec3e75de741f5b53e1/README.md","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/kevinpita/pi-jev-context/blob/640ffa74c376fe8e487e4fec3e75de741f5b53e1/src/jev.ts","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/kevinpita/pi-jev-context/blob/640ffa74c376fe8e487e4fec3e75de741f5b53e1/src/index.ts","note":"固定版本说明与源码；本站未执行该项目。"}],"pinned":false,"topics":["context-pruning","jev","pi","pi-extension","pi-package","typesafe","typescript"]},{"id":"colliber:duckdb-jev","name":"duckdb-jev","author":"colliber","url":"https://github.com/colliber/duckdb-jev","category":"Data & Search","plainSummary":"一个 DuckDB 扩展，在 SQL 中调用 Jev，并把回答转换成 ENUM、数值或 STRUCT 等类型。","plainSummaryEn":"A DuckDB extension that calls Jev from SQL and returns answers as ENUM, numeric, or STRUCT types.","jevDecisionPoint":"对行文本提出 Choice、Score、Noul 问题，按问题定义生成对应 SQL 结果。","jevDecisionPointEn":"Asks Choice, Score, and Noul questions about row text and maps answers to SQL types defined by the questions.","highlightBenefit":"在查询表格或 Parquet 数据时直接使用结构化判断。","highlightBenefitEn":"Uses structured judgments while querying tables or Parquet data.","tags":["database-vector","typed-decisions"],"stars":19,"forks":1,"openIssues":2,"license":"MIT","language":"C++","lastCommitAt":"2026-09-18T08:40:39Z","createdAt":"2026-09-17T12:48:22Z","metadataFetchedAt":"2026-09-21T22:05:36.008Z","avatarUrl":"https://avatars.githubusercontent.com/u/262973864?v=4","summarySource":"human-reviewed","claimStatus":"需要配置外部 Jev API；SQL 类型约束不等于答案正确或隐私无风险。","claimStatusEn":"Requires an external Jev API configuration. SQL type constraints do not establish answer correctness or eliminate privacy risks.","plainSummaryJa":"SQL から Jev を呼び、ENUM、数値、STRUCT などで答えを返す DuckDB 拡張。","plainSummaryKo":"SQL에서 Jev를 호출하고 ENUM, 숫자, STRUCT 등으로 답을 반환하는 DuckDB 확장.","jevDecisionPointJa":"行のテキストを Choice、Score、Noul で評価し、定義に合う SQL 型へ変換する。","jevDecisionPointKo":"행 텍스트를 Choice, Score, Noul로 평가해 정의에 맞는 SQL 타입으로 변환한다.","highlightBenefitJa":"表や Parquet の問い合わせ中に構造化判断を使える。","highlightBenefitKo":"테이블이나 Parquet를 조회하면서 구조화된 판단을 사용할 수 있다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"confirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/colliber/duckdb-jev/blob/fad67eb4a1def91fcf1efbc96bdf5e6e11208644/src/jev_client.cpp","note":"Fixed-commit source reviewed / 已核对固定版本源码"},{"url":"https://github.com/colliber/duckdb-jev/blob/fad67eb4a1def91fcf1efbc96bdf5e6e11208644/README.md","note":"Fixed-commit source reviewed / 已核对固定版本源码"}],"pinned":false},{"id":"choxos:jev-reviewer","name":"jev-reviewer","author":"choxos","url":"https://github.com/choxos/jev-reviewer","category":"Domain & Vertical Tools","plainSummary":"为系统综述从论文与补充材料中挑出原文证据，供研究者逐条核对并导出提取表。","plainSummaryEn":"A systematic-review tool that selects verbatim evidence from papers and supplements for human checking and export.","jevDecisionPoint":"Jev 选择能回答问题的候选行号，代码复制原文并保留文件和页码等位置。","jevDecisionPointEn":"Jev selects candidate line IDs; code copies the original text with file and location references.","highlightBenefit":"把引用、来源位置和人工确认状态对应起来，方便回查。","highlightBenefitEn":"Connects quotes, source locations and human verification status.","tags":["domain-workflows","search-retrieval"],"stars":31,"forks":6,"openIssues":1,"license":"MIT","language":"JavaScript","lastCommitAt":"2026-09-19T20:44:55Z","createdAt":"2026-09-18T14:47:00Z","metadataFetchedAt":"2026-09-21T22:05:38.993Z","avatarUrl":"https://avatars.githubusercontent.com/u/6255852?v=4","summarySource":"human-reviewed","claimStatus":"文件保存在浏览器，但参与判断的文本会经请求送往 Jev；原文引用不保证相关性或医学结论正确。","claimStatusEn":"Files are stored in the browser, but text used for judgment is sent to Jev. Verbatim quotes do not guarantee relevance or medical correctness.","plainSummaryJa":"論文と補足資料から原文の証拠を選び、人が確認して抽出表へ出力するレビュー支援ツール。","plainSummaryKo":"논문·보충 자료에서 원문 증거를 골라 사람이 확인하고 추출 표로 내보내는 리뷰 도구.","jevDecisionPointJa":"Jev が候補行 ID を選び、コードが原文とファイル・位置をコピーする。","jevDecisionPointKo":"Jev가 후보 행 ID를 선택하고 코드가 원문과 파일·위치를 복사한다.","highlightBenefitJa":"引用・出典位置・人による確認状況を対応付ける。","highlightBenefitKo":"인용·출처 위치·사람의 확인 상태를 연결한다.","claimStatusJa":"固定バージョンの文書とソースを確認。実行・性能・安全性は当サイトで独立検証していません。","claimStatusKo":"고정 버전 문서와 소스를 확인했습니다. 실행·성능·안전성은 이 사이트에서 독립적으로 검증하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/choxos/jev-reviewer/blob/b00d05fc888973b0147383dd7315318280d01635/README.md","note":"README.md"},{"url":"https://github.com/choxos/jev-reviewer/blob/b00d05fc888973b0147383dd7315318280d01635/server.js","note":"server.js"}],"pinned":false,"topics":["clinical-trials","cochrane","data-extraction","endnote","evidence-based-medicine","evidence-synthesis","jev","literature-review","local-first","meta-analysis","open-science","pdf","prisma","research-tools","risk-of-bias","rob2","systematic-review","typesafe","voice","zotero"]},{"id":"eugeneboondock:jevsql","name":"jevsql","author":"EugeneBoondock","url":"https://github.com/EugeneBoondock/jevsql","category":"Data & Search","plainSummary":"给 SQLite 加入 Jev 语义判断，可过滤、排序、匹配记录并追踪判断所用的证据。","plainSummaryEn":"Adds Jev semantic judgments to SQLite for filtering, ranking, matching and tracking decision evidence.","jevDecisionPoint":"把行数据和候选问题发给 Jev，再把答案映射为 SQL 可查询结果。","jevDecisionPointEn":"Sends row data and typed questions to Jev and maps answers into queryable SQL results.","highlightBenefit":"提供批处理、缓存和预算控制，并保留判断历史。","highlightBenefitEn":"Provides batching, caching, budgets and decision history.","tags":["database-vector","search-retrieval","classification-ranking"],"stars":3,"forks":1,"openIssues":0,"license":"MIT","language":"JavaScript","lastCommitAt":"2026-09-19T23:34:05Z","createdAt":"2026-09-18T00:24:50Z","metadataFetchedAt":"2026-09-21T22:05:41.985Z","avatarUrl":"https://avatars.githubusercontent.com/u/80263867?v=4","summarySource":"human-reviewed","claimStatus":"已核对固定版本 README 与集成源码。","claimStatusEn":"README and integration source reviewed at a fixed commit; not independently run or benchmarked by this site.","plainSummaryJa":"SQLite に Jev の意味判断を加え、絞り込み、順位付け、照合と根拠追跡を行う。","plainSummaryKo":"SQLite에 Jev 의미 판단을 더해 필터링, 순위, 매칭과 근거 추적을 수행한다.","jevDecisionPointJa":"行データと質問を Jev に送り、回答を SQL で扱える結果へ変換する。","jevDecisionPointKo":"행 데이터와 질문을 Jev에 보내 답변을 SQL 조회 결과로 바꾼다.","highlightBenefitJa":"一括処理、キャッシュ、予算管理、判断履歴を備える。","highlightBenefitKo":"일괄 처리, 캐시, 예산 제어와 판단 기록을 제공한다.","claimStatusJa":"固定版の README と実装を確認済み。このサイトでは独立した実行・性能検証をしていません。","claimStatusKo":"고정 버전 README와 구현을 확인했습니다. 이 사이트에서 독립적인 실행·성능 검증은 하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/EugeneBoondock/jevsql/blob/e0b42f3b4312189f0b9c3996818675c714f126e7/src/client.mjs#L5","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/EugeneBoondock/jevsql/blob/e0b42f3b4312189f0b9c3996818675c714f126e7/README.md","note":"固定版本的实现与说明 / Pinned source and documentation"}],"pinned":false,"topics":["ai","jev","llm","semantic-search","sql","sqlite","typesafe"]},{"id":"n3ndor:n8n-nodes-typesafe-jev","name":"n8n-nodes-typesafe-jev","author":"n3ndor","url":"https://github.com/n3ndor/n8n-nodes-typesafe-jev","category":"MCP & Integrations","plainSummary":"n8n 的社区 TypeSafe Jev 节点，让工作流提交结构化判断题。","plainSummaryEn":"An n8n community node for submitting typed questions to TypeSafe Jev.","jevDecisionPoint":"从输入 item 构造状态和问题，调用 Jev，再附加或单独输出答案。","jevDecisionPointEn":"Builds state and questions from input items, calls Jev and appends or separately outputs answers.","highlightBenefit":"可用表单或 JSON 配置问题，也可把节点作为 Agent 工具。","highlightBenefitEn":"Supports form or JSON question configuration and use as an Agent tool.","tags":["mcp-integrations","typed-decisions"],"stars":2,"forks":1,"openIssues":2,"license":"MIT","language":"TypeScript","lastCommitAt":"2026-09-19T13:54:29Z","createdAt":"2026-09-18T17:10:29Z","metadataFetchedAt":"2026-09-21T22:05:44.989Z","avatarUrl":"https://avatars.githubusercontent.com/u/107216827?v=4","summarySource":"human-reviewed","claimStatus":"已核对固定版本 README 与调用源码。","claimStatusEn":"Reviewed the fixed-version README and call path.","plainSummaryJa":"TypeSafe Jev に型付き質問を送る n8n のコミュニティノード。","plainSummaryKo":"TypeSafe Jev에 구조화된 질문을 보내는 n8n 커뮤니티 노드.","jevDecisionPointJa":"入力 item から状態と質問を作り、Jev の回答を追記または単独で出力する。","jevDecisionPointKo":"입력 item에서 상태·질문을 만들고 Jev 응답을 추가하거나 별도로 출력한다.","highlightBenefitJa":"フォーム・JSON で質問を設定でき、Agent ツールとしても利用できる。","highlightBenefitKo":"양식·JSON으로 질문을 설정하고 Agent 도구로도 사용할 수 있다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/n3ndor/n8n-nodes-typesafe-jev/blob/22ae978a36149982fe14b8556eb0bab0250c2450/README.md","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/n3ndor/n8n-nodes-typesafe-jev/blob/22ae978a36149982fe14b8556eb0bab0250c2450/nodes/TypeSafeJev/transport.ts","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/n3ndor/n8n-nodes-typesafe-jev/blob/22ae978a36149982fe14b8556eb0bab0250c2450/nodes/TypeSafeJev/TypeSafeJev.node.ts","note":"固定版本说明与源码；本站未执行该项目。"}],"pinned":false},{"id":"kyotofin:tax-doc-classifier","name":"tax-doc-classifier","author":"kyotofin","url":"https://github.com/kyotofin/tax-doc-classifier","category":"Domain & Vertical Tools","plainSummary":"税务文档页面分类器，用 Jev 从预先定义的 IRS 表格与页面类型中选择类别。","plainSummaryEn":"A tax-document page classifier that uses Jev to choose from predefined IRS form and page categories.","jevDecisionPoint":"提取 PDF 页面的文本后发送给 Jev，返回表格类型、页面类型及置信度。","jevDecisionPointEn":"Extracts PDF page text and sends it to Jev for form type, page type, and confidence.","highlightBenefit":"将固定表格目录与每页分类结果连接起来，供后续文档流程使用。","highlightBenefitEn":"Connects a fixed form catalog to per-page classifications for downstream document workflows.","tags":["domain-workflows","classification-ranking"],"stars":339,"forks":35,"openIssues":0,"license":"Apache-2.0","language":"TypeScript","lastCommitAt":"2026-09-20T10:34:44Z","createdAt":"2026-09-18T14:40:45Z","metadataFetchedAt":"2026-09-21T22:05:48.001Z","avatarUrl":"https://avatars.githubusercontent.com/u/291348760?v=4","summarySource":"human-reviewed","claimStatus":"“100%”及单页成本是作者针对其语料的自报结果，不是所有税务文档的准确率保证。","claimStatusEn":"The 100% figure and per-page cost are author-reported results on their own corpus, not an accuracy guarantee for all tax documents. This site has not retested them or verified tax judgment capabilities.","plainSummaryJa":"事前定義した IRS の書式とページ種別を Jev が選ぶ税務書類分類器。","plainSummaryKo":"미리 정의된 IRS 양식과 페이지 유형을 Jev가 선택하는 세무 문서 분류기.","jevDecisionPointJa":"PDF ページのテキストを抽出し、書式、ページ種別、確信度を取得する。","jevDecisionPointKo":"PDF 페이지 텍스트를 추출해 양식, 페이지 유형, 신뢰도를 얻는다.","highlightBenefitJa":"固定の書式一覧と各ページの分類を後続の処理につなぐ。","highlightBenefitKo":"고정 양식 목록과 각 페이지 분류를 후속 처리에 연결한다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"confirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/kyotofin/tax-doc-classifier/blob/6afcf701395466d7c936ec8178daf017b9d96b0c/src/backend.ts","note":"Fixed-commit source reviewed / 已核对固定版本源码"},{"url":"https://github.com/kyotofin/tax-doc-classifier/blob/6afcf701395466d7c936ec8178daf017b9d96b0c/README.md","note":"Fixed-commit source reviewed / 已核对固定版本源码"}],"pinned":false},{"id":"dabit3:jev-experiments","name":"jev-experiments","author":"dabit3","url":"https://github.com/dabit3/jev-experiments","category":"Security & Guardrails","plainSummary":"一组 Jev 开发工具实验，其中 Commit Sentry 对暂存 diff 的片段进行语义风险检查。","plainSummaryEn":"A collection of Jev developer-tool experiments, including Commit Sentry for semantic checks on staged diff hunks.","jevDecisionPoint":"逐片段判断密钥泄漏、破坏性变更等风险，由本地规则决定警告或阻止提交。","jevDecisionPointEn":"Judges risks such as exposed credentials or destructive changes; local rules warn or block a commit.","highlightBenefit":"把问题标签与对应 diff 放在提交前，便于开发者检查。","highlightBenefitEn":"Presents risk labels with the corresponding diff before commit.","tags":["cli-git-gates","security-guardrails","coding-agents"],"stars":355,"forks":26,"openIssues":18,"license":null,"language":"TypeScript","lastCommitAt":"2026-09-20T02:00:47Z","createdAt":"2026-09-17T23:36:37Z","metadataFetchedAt":"2026-09-21T22:05:51.003Z","avatarUrl":"https://avatars.githubusercontent.com/u/1857282?v=4","summarySource":"human-reviewed","claimStatus":"仓库说明由 Devin 构建，性能表是作者特定运行记录；不能替代确定性密钥扫描或完整代码审查。","claimStatusEn":"The repository says the demos were built by Devin; performance tables describe author-specific runs. Not a replacement for deterministic secret scanning or full code review, and not retested here.","plainSummaryJa":"ステージ済み diff を意味的に確認する Commit Sentry など、Jev 開発ツールの実験集。","plainSummaryKo":"스테이징 diff를 의미적으로 검사하는 Commit Sentry 등 Jev 개발 도구 실험 모음.","jevDecisionPointJa":"秘密情報の露出や破壊的変更を評価し、ローカル規則が警告・停止する。","jevDecisionPointKo":"자격 정보 노출이나 파괴적 변경을 평가하고 로컬 규칙이 경고·중단한다.","highlightBenefitJa":"コミット前にリスクラベルと該当 diff を確認できる。","highlightBenefitKo":"커밋 전에 위험 라벨과 해당 diff를 확인한다.","claimStatusJa":"固定バージョンの文書とソースを確認。実行・性能・安全性は当サイトで独立検証していません。","claimStatusKo":"고정 버전 문서와 소스를 확인했습니다. 실행·성능·안전성은 이 사이트에서 독립적으로 검증하지 않았습니다.","licenseStatus":"unconfirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/dabit3/jev-experiments/blob/ed07e87c24822b9258dfd95c63e3d5f436d95b1a/README.md","note":"README.md"},{"url":"https://github.com/dabit3/jev-experiments/blob/ed07e87c24822b9258dfd95c63e3d5f436d95b1a/commit-sentry/src/jev.ts","note":"commit-sentry/src/jev.ts"},{"url":"https://github.com/dabit3/jev-experiments/blob/ed07e87c24822b9258dfd95c63e3d5f436d95b1a/commit-sentry/README.md","note":"commit-sentry/README.md"}],"pinned":false},{"id":"skyvern-ai:jevscape","name":"jevscape","author":"Skyvern-AI","url":"https://github.com/Skyvern-AI/jevscape","category":"High-Frequency & Simulation","plainSummary":"RuneBench 的 Jev 扩展，使用 rs-sdk 的有界动作目录驱动 RuneScape 游戏任务。","plainSummaryEn":"A RuneBench extension using Jev and a bounded rs-sdk action catalog for RuneScape tasks.","jevDecisionPoint":"按游戏状态选择目标动作、当前 tick 干预与下一次询问间隔。","jevDecisionPointEn":"Chooses goal actions, tick interventions and the next polling interval from game state.","highlightBenefit":"提供动作分布面板、运行记录和 burst/tick 两种控制方式。","highlightBenefitEn":"Provides action-distribution views, run logs and burst or tick control modes.","tags":["games-simulation","typed-decisions"],"stars":7,"forks":0,"openIssues":1,"license":null,"language":"TypeScript","lastCommitAt":"2026-09-18T19:13:24Z","createdAt":"2026-09-18T18:10:42Z","metadataFetchedAt":"2026-09-21T22:05:54.045Z","avatarUrl":"https://avatars.githubusercontent.com/u/141457985?v=4","summarySource":"human-reviewed","claimStatus":"项目有独立 Jev 适配器，不是上游零改动副本；演示与成本结果不代表完整任务基准胜出。 已核对固定版本 README 与集成源码。","claimStatusEn":"This repository has its own Jev adapter rather than being an unchanged upstream copy. Demos and cost figures do not establish comprehensive benchmark superiority. README and integration source reviewed at a fixed commit; not independently run or benchmarked by this site.","plainSummaryJa":"rs-sdk の限定操作候補と Jev で RuneScape タスクを動かす RuneBench 拡張。","plainSummaryKo":"rs-sdk의 제한된 동작 목록과 Jev로 RuneScape 작업을 수행하는 RuneBench 확장이다.","jevDecisionPointJa":"状態から目標動作、tick 内の介入、次の問い合わせ間隔を選ぶ。","jevDecisionPointKo":"게임 상태로 목표 동작, tick 개입과 다음 질의 간격을 고른다.","highlightBenefitJa":"動作分布画面、実行記録、burst/tick 制御を提供する。","highlightBenefitKo":"동작 분포 화면, 실행 기록과 burst·tick 제어를 제공한다.","claimStatusJa":"固定版の README と実装を確認済み。このサイトでは独立した実行・性能検証をしていません。","claimStatusKo":"고정 버전 README와 구현을 확인했습니다. 이 사이트에서 독립적인 실행·성능 검증은 하지 않았습니다.","licenseStatus":"unconfirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/Skyvern-AI/jevscape/blob/8fe4d37349fc6a8d03ac1c918302eef254bb56c3/agents/jev/jev-client.ts#L3","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/Skyvern-AI/jevscape/blob/8fe4d37349fc6a8d03ac1c918302eef254bb56c3/README.md","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/Skyvern-AI/jevscape/blob/8fe4d37349fc6a8d03ac1c918302eef254bb56c3/agents/jev/README.md","note":"固定版本的实现与说明 / Pinned source and documentation"}],"pinned":false},{"id":"monteduro:killmyidea","name":"killmyidea","author":"monteduro","url":"https://github.com/monteduro/killmyidea","category":"Decision Tools","plainSummary":"创业点子评估演示，用 Jev 的多项评分给出 KILL、FIX 或 SHIP 标签。","plainSummaryEn":"A startup-idea evaluation demo assigning KILL, FIX or SHIP labels from Jev scores.","jevDecisionPoint":"Jev 回答评分、分类和清晰度问题，本地加权与门槛计算最终标签。","jevDecisionPointEn":"Jev answers scoring, category and clarity questions; local weights and gates calculate the final label.","highlightBenefit":"展示结构化评估过程，不是市场验证、商业成功预测或投资建议。","highlightBenefitEn":"Shows a structured assessment process, not market validation, a success forecast or investment advice.","tags":["domain-workflows","typed-decisions"],"stars":67,"forks":7,"openIssues":0,"license":null,"language":"TypeScript","lastCommitAt":"2026-09-18T10:56:36Z","createdAt":"2026-09-17T11:08:29Z","metadataFetchedAt":"2026-09-21T22:05:57.016Z","avatarUrl":"https://avatars.githubusercontent.com/u/5638815?v=4","summarySource":"human-reviewed","claimStatus":"标签来自模型评分和本地加权规则，不是市场或商业可行性的独立验证。","claimStatusEn":"Labels derive from model scores and local weights, not independent market or business validation.","plainSummaryJa":"Jev の複数スコアから KILL・FIX・SHIP を付ける起業アイデア評価デモ。","plainSummaryKo":"Jev의 여러 점수로 KILL·FIX·SHIP 라벨을 부여하는 창업 아이디어 평가 데모.","jevDecisionPointJa":"採点・分類・明瞭さを Jev に問い、ローカルの重みとゲートで最終ラベルを計算する。","jevDecisionPointKo":"채점·분류·명확성을 Jev에 묻고 로컬 가중치·조건으로 최종 라벨을 계산한다.","highlightBenefitJa":"評価手順の例であり、市場検証・成功予測・投資助言ではない。","highlightBenefitKo":"평가 과정 예제이며 시장 검증·성공 예측·투자 조언은 아니다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"unconfirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/monteduro/killmyidea/blob/bc853421f2eb6f17da435621896dd6cd881a051c/README.md","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/monteduro/killmyidea/blob/bc853421f2eb6f17da435621896dd6cd881a051c/src/lib/typesafe.ts","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/monteduro/killmyidea/blob/bc853421f2eb6f17da435621896dd6cd881a051c/src/lib/scoring.ts","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/monteduro/killmyidea/blob/bc853421f2eb6f17da435621896dd6cd881a051c/api/evaluate.ts","note":"固定版本说明与源码；本站未执行该项目。"}],"pinned":false},{"id":"rafalwilinski:vibecheck","name":"vibecheck","author":"RafalWilinski","url":"https://github.com/RafalWilinski/vibecheck","category":"Creative Tools","plainSummary":"在 X 发帖前显示一个 Jev 评分卡，检查草稿的清晰度、语气、冒犯倾向等维度。","plainSummaryEn":"Shows a Jev scorecard before posting on X, assessing draft clarity, tone, and possible offensiveness.","jevDecisionPoint":"把草稿及回复/引用上下文交给 Jev，返回多个评分与发帖建议。","jevDecisionPointEn":"Sends the draft and reply or quote context to Jev for multiple scores and a posting suggestion.","highlightBenefit":"在发送前集中查看对文案的不同评价角度。","highlightBenefitEn":"Collects several perspectives on a draft before it is sent.","tags":["browser-automation","creative-multimedia","evaluation-benchmarks"],"stars":46,"forks":5,"openIssues":1,"license":null,"language":"JavaScript","lastCommitAt":"2026-09-20T16:24:38Z","createdAt":"2026-09-18T14:41:49Z","metadataFetchedAt":"2026-09-21T22:06:00.034Z","avatarUrl":"https://avatars.githubusercontent.com/u/3391616?v=4","summarySource":"human-reviewed","claimStatus":"传播性、后悔风险等是模型主观估计，不保证实际反应或发帖安全；图片描述需可选的外部视觉模型，本站未验证效果。","claimStatusEn":"Virality and regret risk are subjective model estimates, not guarantees of audience response or posting safety. Image descriptions require an optional external vision model; this site has not verified effectiveness.","plainSummaryJa":"X への投稿前に、明瞭さ、語調、不快さなどを Jev で採点するカードを表示する。","plainSummaryKo":"X에 게시하기 전에 명확성, 어조, 불쾌감 등을 Jev로 평가한 카드를 보여준다.","jevDecisionPointJa":"下書きと返信・引用の文脈を送り、複数の評価と投稿の提案を受け取る。","jevDecisionPointKo":"초안과 답글·인용 맥락을 보내 여러 평가와 게시 제안을 받는다.","highlightBenefitJa":"送信前に文章を複数の観点から見直せる。","highlightBenefitKo":"전송 전에 글을 여러 관점에서 검토할 수 있다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"unconfirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/RafalWilinski/vibecheck/blob/badf9dad5deecababd8afe43a574a5ea6d376711/background.js","note":"Fixed-commit source reviewed / 已核对固定版本源码"},{"url":"https://github.com/RafalWilinski/vibecheck/blob/badf9dad5deecababd8afe43a574a5ea6d376711/README.md","note":"Fixed-commit source reviewed / 已核对固定版本源码"}],"pinned":false},{"id":"rinnecoder:jev-behavior-study","name":"jev-behavior-study","author":"RINNECODER","url":"https://github.com/RINNECODER/jev-behavior-study","category":"Evaluation & Observability","plainSummary":"针对 Jev 1.13.0 的独立行为研究，记录不同问题表述、输入条件及游戏任务下的成功与失败。","plainSummaryEn":"An independent Jev 1.13.0 behavior study recording successes and failures across question framing, input conditions and games.","jevDecisionPoint":"向固定任务发送受控变体，保存选择结果、概率和原始请求响应。","jevDecisionPointEn":"Sends controlled variants of fixed tasks and records choices, probabilities and raw request-response evidence.","highlightBenefit":"让读者按具体案例查看证据，区分小任务通过与复杂任务能力。","highlightBenefitEn":"Lets readers inspect individual cases rather than infer broad capability from simple-task success.","tags":["evaluation-benchmarks","games-simulation"],"stars":3,"forks":0,"openIssues":1,"license":"MIT","language":"Python","lastCommitAt":"2026-09-17T05:06:01Z","createdAt":"2026-09-16T08:26:33Z","metadataFetchedAt":"2026-09-21T22:06:03.005Z","avatarUrl":"https://avatars.githubusercontent.com/u/93304526?v=4","summarySource":"human-reviewed","claimStatus":"作者的合成、任务特定实验；重复调用不等于独立新题，也不是官方基准或同行评审结论。本站未重跑。","claimStatusEn":"Author-run synthetic, task-specific experiments; repeated calls are not independent problems, and this is not an official benchmark or a peer-reviewed conclusion. Not rerun here.","plainSummaryJa":"質問表現・入力条件・ゲーム課題における成功と失敗を記録する Jev 1.13.0 の独立研究。","plainSummaryKo":"질문 표현·입력 조건·게임 과제의 성공과 실패를 기록한 Jev 1.13.0 독립 연구.","jevDecisionPointJa":"固定課題の条件を変え、選択・確率・要求と応答の原記録を残す。","jevDecisionPointKo":"고정 과제의 조건을 바꾸고 선택·확률·요청과 응답 원문을 기록한다.","highlightBenefitJa":"簡単な課題の成功を一般化せず、個々の事例を確認できる。","highlightBenefitKo":"단순 과제의 성공을 일반화하지 않고 개별 사례를 확인한다.","claimStatusJa":"固定バージョンの文書とソースを確認。実行・性能・安全性は当サイトで独立検証していません。","claimStatusKo":"고정 버전 문서와 소스를 확인했습니다. 실행·성능·안전성은 이 사이트에서 독립적으로 검증하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/RINNECODER/jev-behavior-study/blob/e4a1d7ec691a91f33d3b5879a780e6f27328f173/README.md","note":"README.md"},{"url":"https://github.com/RINNECODER/jev-behavior-study/blob/e4a1d7ec691a91f33d3b5879a780e6f27328f173/behavior_study.py","note":"behavior_study.py"}],"pinned":false},{"id":"4esv:jev-eval","name":"jev-eval","author":"4esv","url":"https://github.com/4esv/jev-eval","category":"Evaluation & Observability","plainSummary":"在有标签的分类任务上比较 Jev 与 OpenRouter 模型的准确率、校准、时延和成本。","plainSummaryEn":"Compares Jev and OpenRouter models on labeled tasks for accuracy, calibration, latency and cost.","jevDecisionPoint":"对相同任务收集模型判断，并计算置信区间及重复输入的稳定性。","jevDecisionPointEn":"Collects judgments on matched tasks and computes confidence intervals and repeat-input stability.","highlightBenefit":"发布数据处理、调用与统计代码，以及特定模型的结果。","highlightBenefitEn":"Publishes data processing, runner and statistics code with model-specific results.","tags":["evaluation-benchmarks","classification-ranking"],"stars":1,"forks":0,"openIssues":0,"license":null,"language":"Python","lastCommitAt":"2026-09-21T21:30:29Z","createdAt":"2026-09-18T02:14:18Z","metadataFetchedAt":"2026-09-21T22:06:06.032Z","avatarUrl":"https://avatars.githubusercontent.com/u/20231779?v=4","summarySource":"human-reviewed","claimStatus":"公开比较为每个任务 300 项、指定模型与日期；优势因任务变化，不应写成全领域领先或无偏证明。 已核对固定版本 README 与集成源码。","claimStatusEn":"The published comparison uses 300 items per task with specified models and dates. Advantages vary by task and do not demonstrate universal superiority or absence of bias. README and integration source reviewed at a fixed commit; not independently run or benchmarked by this site.","plainSummaryJa":"正解付き分類課題で Jev と OpenRouter モデルの正確さ、較正、遅延、費用を比較する。","plainSummaryKo":"정답이 있는 분류 과제로 Jev와 OpenRouter 모델의 정확도, 보정, 지연과 비용을 비교한다.","jevDecisionPointJa":"同じ課題の判断を集め、信頼区間と同一入力の安定性を計算する。","jevDecisionPointKo":"같은 과제의 판단을 모아 신뢰구간과 반복 입력 안정성을 계산한다.","highlightBenefitJa":"前処理、呼び出し、統計コードとモデル別結果を公開する。","highlightBenefitKo":"전처리, 호출, 통계 코드와 모델별 결과를 공개한다.","claimStatusJa":"固定版の README と実装を確認済み。このサイトでは独立した実行・性能検証をしていません。","claimStatusKo":"고정 버전 README와 구현을 확인했습니다. 이 사이트에서 독립적인 실행·성능 검증은 하지 않았습니다.","licenseStatus":"unconfirmed","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/4esv/jev-eval/blob/dcfd6db29dfee2a107ff6e018176c6577002f887/evaljev/runners.py#L15","note":"固定版本的实现与说明 / Pinned source and documentation"},{"url":"https://github.com/4esv/jev-eval/blob/dcfd6db29dfee2a107ff6e018176c6577002f887/README.md","note":"固定版本的实现与说明 / Pinned source and documentation"}],"pinned":false},{"id":"tonyzdev:pij","name":"PiJ","author":"tonyzdev","url":"https://github.com/tonyzdev/PiJ","category":"Codebase & Graph Pathfinding","plainSummary":"基于 Pi 的终端编码 Agent，主模型负责推理、改代码与工具调用，Jev 提供辅助判断。","plainSummaryEn":"A Pi-based terminal coding Agent whose main model handles reasoning, edits and tools while Jev provides advice.","jevDecisionPoint":"Jev 推荐 skill、重排真实源码候选并分类工具失败；不自动重试或批准权限。","jevDecisionPointEn":"Jev suggests skills, reranks actual source candidates and triages failures without automatic retries or permission approval.","highlightBenefit":"保留原始路径、行号、源码与错误输出；作者的有限实验不代表普遍收益。","highlightBenefitEn":"Preserves paths, line numbers, source and errors; the author’s bounded experiments do not establish universal gains.","tags":["coding-agents","typed-decisions"],"stars":4,"forks":0,"openIssues":0,"license":"MIT","language":"TypeScript","lastCommitAt":"2026-09-20T02:29:17Z","createdAt":"2026-09-17T19:31:31Z","metadataFetchedAt":"2026-09-21T22:06:12.030Z","avatarUrl":"https://avatars.githubusercontent.com/u/69023709?v=4","summarySource":"human-reviewed","claimStatus":"作者提供有限任务实验；本站核对实现与证据边界，未独立复测解题率、成本或耗时。","claimStatusEn":"The author provides bounded task experiments; this site reviewed implementation and evidence limits without independently retesting resolve rate, cost or latency.","plainSummaryJa":"主モデルが推論・編集・ツール実行を担い、Jev が補助判断する Pi ベースのターミナル Agent。","plainSummaryKo":"주 모델이 추론·수정·도구 실행을 맡고 Jev가 보조 판단하는 Pi 기반 터미널 Agent.","jevDecisionPointJa":"skill 提案・実在するソース候補の再順位付け・失敗分類を行い、自動再試行や権限承認はしない。","jevDecisionPointKo":"skill 제안·실제 소스 후보 재정렬·실패 분류를 수행하며 자동 재시도나 권한 승인은 하지 않는다.","highlightBenefitJa":"パス・行番号・ソース・エラーを保持。作者の限定実験は一般的効果の保証ではない。","highlightBenefitKo":"경로·줄 번호·소스·오류를 보존하며 저자의 제한된 실험은 일반적 효과를 보장하지 않는다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T06:53:39.508919+00:00","evidence":[{"url":"https://github.com/tonyzdev/PiJ/blob/9588d6078af3f3be2b52906e0968ccf8e1542d74/README.md","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/tonyzdev/PiJ/blob/9588d6078af3f3be2b52906e0968ccf8e1542d74/README.zh-CN.md","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/tonyzdev/PiJ/blob/9588d6078af3f3be2b52906e0968ccf8e1542d74/src/jev.ts","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/tonyzdev/PiJ/blob/9588d6078af3f3be2b52906e0968ccf8e1542d74/src/gateway.ts","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/tonyzdev/PiJ/blob/9588d6078af3f3be2b52906e0968ccf8e1542d74/src/decisions.ts","note":"固定版本说明与源码；本站未执行该项目。"},{"url":"https://github.com/tonyzdev/PiJ/blob/9588d6078af3f3be2b52906e0968ccf8e1542d74/src/extension.ts","note":"固定版本说明与源码；本站未执行该项目。"}],"pinned":false,"topics":["ai-agent","coding-agent","jev","pi","terminal","typescript","vercel-ai-gateway"]},{"id":"valentynkit:jev-commit","name":"jev-commit","author":"valentynkit","url":"https://github.com/valentynkit/jev-commit","category":"Decision Tools","plainSummary":"commit-msg 钩子：一次 Jev 请求对照暂存 diff 给提交说明打五个 Noul，默认只警告。默认拦住提交的是本地正则腰带对新增行的高置信命中；secret_shaped Noul 只在 --strict 时参与拦截。","plainSummaryEn":"A commit-msg hook that scores the message against the staged diff with one Jev request of five Nouls and warns by default. Default blocking is a local regex belt on high-precision added-line hits; the secret_shaped Noul blocks only with --strict.","jevDecisionPoint":"一次请求里五个 Noul：说明是否可核对、是否与 hunk 相符、是否留下调试代码、是否有未提及改动、新增行是否像密钥。默认拦截来自正则腰带；secret_shaped 只在 --strict 下参与拦截。","jevDecisionPointEn":"Five Noul questions in one request: checkable message, hunk match, debug leftovers, unmentioned work, and credential-shaped added lines. Default blocking is the regex belt; secret_shaped participates only under --strict.","highlightBenefit":"把提交说明核对做成可设阈值的概率，而不是再读一段模型评语；接口失败时仍放行提交。","highlightBenefitEn":"Turns commit-message review into thresholded probabilities instead of prose; API failures still allow the commit.","tags":["typed-decisions","classification-ranking"],"stars":9,"forks":0,"openIssues":1,"license":"MIT","lastCommitAt":"2026-09-19T07:47:00Z","createdAt":"2026-09-18T20:11:08Z","metadataFetchedAt":"2026-09-21T22:06:15.069Z","avatarUrl":"https://avatars.githubusercontent.com/u/35431524?v=4","summarySource":"curated","claimStatus":"优先保留投稿者与仓库原文，缺失语言自动补充。自动检查已确认 Jev 集成证据。","claimStatusEn":"Author and repository text is preserved where clear; missing languages are enriched automatically. Checks establish source-level Jev integration, not runtime, safety, or performance validation.","plainSummaryJa":"commit-msg フック。1 回の Jev リクエストで staged diff と照合し 5 つの Noul を付け、既定では警告のみ。既定の停止は追加行に対するローカル正規表現ベルトの高信頼ヒット。secret_shaped Noul は --strict のときだけ止める。","plainSummaryKo":"commit-msg 훅. Jev 요청 한 번으로 staged diff와 대조해 Noul 다섯 개를 매기고 기본은 경고만 한다. 기본으로 커밋을 막는 것은 추가 줄에 대한 로컬 정규식 벨트의 고정밀 적중이다. secret_shaped Noul은 --strict일 때만 막는다.","jevDecisionPointJa":"1 リクエストで 5 つの Noul。メッセージが照合可能か、hunk と一致するか、デバッグ残骸があるか、未言及の変更があるか、追加行が認証情報らしいか。既定の停止は正規表現ベルト。secret_shaped は --strict のときだけ。","jevDecisionPointKo":"한 요청에서 Noul 다섯 개. 메시지가 대조 가능한지, hunk와 맞는지, 디버그 잔여물이 있는지, 언급되지 않은 변경이 있는지, 추가 줄이 자격 증명처럼 보이는지. 기본 차단은 정규식 벨트다. secret_shaped는 --strict에서만 참여한다.","highlightBenefitJa":"コミットメッセージの照合をしきい値付き確率にし、モデルの評論文を読まなくてよい。API 失敗時もコミットは通す。","highlightBenefitKo":"커밋 메시지 대조를 임계값 있는 확률로 만들어 모델 평문을 읽지 않아도 된다. API가 실패해도 커밋은 통과한다.","evidence":[{"url":"https://github.com/valentynkit/jev-commit/blob/311e163b8abb9c333132155bc2e0bbfac4f36283/README.md","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/valentynkit/jev-commit/blob/311e163b8abb9c333132155bc2e0bbfac4f36283/jev_commit/jev.py","note":"固定版本的 Jev 集成与说明来源"}]},{"id":"valentynkit:jev-belay","name":"jev-belay","author":"valentynkit","url":"https://github.com/valentynkit/jev-belay","category":"Decision Tools","plainSummary":"Claude Code 的 Stop 钩子：本地先看本轮是否改过文件、是否已有通过的检查，只有这时才问 Jev 结束语是不是未经核实的完成声明。","plainSummaryEn":"A Claude Code Stop hook that reads the turn transcript locally and calls Jev only when files changed and no check has passed, blocking an unverified “done”.","jevDecisionPoint":"四个问题：结束语是否声称完成、是否声称检查已通过、该任务是否值得跑检查，以及 complete / partial / blocked / other。","jevDecisionPointEn":"Four questions: whether the closing message claims done, claims checks passed, whether verification applies, and a complete / partial / blocked / other choice.","highlightBenefit":"有通过的检查就不发请求；出错则放行，避免钩子把会话卡死。","highlightBenefitEn":"Skips the model when a check already passed; every error path lets the turn end.","tags":["typed-decisions","classification-ranking"],"stars":17,"forks":1,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-20T09:06:41Z","createdAt":"2026-09-18T20:11:17Z","metadataFetchedAt":"2026-09-21T22:06:17.990Z","avatarUrl":"https://avatars.githubusercontent.com/u/35431524?v=4","summarySource":"curated","claimStatus":"优先保留投稿者与仓库原文，缺失语言自动补充。自动检查已确认 Jev 集成证据。","claimStatusEn":"Author and repository text is preserved where clear; missing languages are enriched automatically. Checks establish source-level Jev integration, not runtime, safety, or performance validation.","plainSummaryJa":"Claude Code の Stop フック。まずローカルで今ターンにファイル変更があったか、通過した検査があるかを見、そのときだけ Jev に終了文が未確認の完了宣言かを問う。","plainSummaryKo":"Claude Code Stop 훅. 먼저 로컬에서 이번 턴에 파일이 바뀌었는지, 통과한 검사가 있는지를 보고, 그때만 Jev에게 마무리 말이 확인되지 않은 완료 선언인지 묻는다.","jevDecisionPointJa":"4 問。終了文が完了を主張するか、検査通過を主張するか、このタスクに検査が意味を持つか、および complete / partial / blocked / other。","jevDecisionPointKo":"질문 네 개. 마무리 말이 완료를 주장하는지, 검사 통과를 주장하는지, 이 작업에 검사가 의미 있는지, 그리고 complete / partial / blocked / other.","highlightBenefitJa":"通過した検査があればリクエストを出さない。エラー時は通して、フックがセッションを止めない。","highlightBenefitKo":"통과한 검사가 있으면 요청을 보내지 않는다. 오류 시에는 통과시켜 훅이 세션을 멈추지 않게 한다.","evidence":[{"url":"https://github.com/valentynkit/jev-belay/blob/98f39e04492068e39b48962183acf43b4337da32/README.md","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/valentynkit/jev-belay/blob/98f39e04492068e39b48962183acf43b4337da32/tools/fake-jev.mjs","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/valentynkit/jev-belay/blob/98f39e04492068e39b48962183acf43b4337da32/belay.mjs","note":"固定版本的 Jev 集成与说明来源"}]},{"id":"valentynkit:jev.nvim","name":"jev.nvim","author":"valentynkit","url":"https://github.com/valentynkit/jev.nvim","category":"Decision Tools","plainSummary":"Neovim 插件：用自然语言问当前 buffer，Treesitter 按函数切开，Jev 给每个函数打概率；全部命中按概率进入 quickfix。","plainSummaryEn":"A Neovim plugin that asks a plain-language question of every function in the buffer and lists every Jev hit in quickfix, ranked by probability.","jevDecisionPoint":"把同一问题套到每个函数的源码上，返回概率；装不下就拆成多次请求。全部命中写入 quickfix，只有达到阈值的才打 virtual text 标记。","jevDecisionPointEn":"Scores each Treesitter function against the user question, splitting across requests if the buffer does not fit. Every hit is listed in quickfix; only hits at or above the threshold get buffer marks.","highlightBenefit":"用问题而不是正则找“会拼 SQL”这类跨语言形状，并直接接上已有的 quickfix 编辑流。","highlightBenefitEn":"Finds cross-language shapes that grep misses, and lands the ranked hits in the existing quickfix loop.","tags":["typed-decisions","classification-ranking"],"stars":3,"forks":0,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-19T07:52:26Z","createdAt":"2026-09-18T20:11:23Z","metadataFetchedAt":"2026-09-21T22:06:21.029Z","avatarUrl":"https://avatars.githubusercontent.com/u/35431524?v=4","summarySource":"curated","claimStatus":"优先保留投稿者与仓库原文，缺失语言自动补充。自动检查已确认 Jev 集成证据。","claimStatusEn":"Author and repository text is preserved where clear; missing languages are enriched automatically. Checks establish source-level Jev integration, not runtime, safety, or performance validation.","plainSummaryJa":"Neovim プラグイン。自然言語で今の buffer に問い、Treesitter が関数単位に切り、Jev が各関数に確率を付ける。命中はすべて確率順で quickfix に入る。","plainSummaryKo":"Neovim 플러그인. 자연어로 현재 buffer에 묻고, Treesitter가 함수 단위로 나눈 뒤 Jev가 함수마다 확률을 매긴다. 모든 적중은 확률 순으로 quickfix에 올라간다.","jevDecisionPointJa":"同じ質問を各関数のソースに当てて確率を返す。収まらなければリクエストを分割する。命中はすべて quickfix に入り、しきい値以上だけ virtual text の印が付く。","jevDecisionPointKo":"같은 질문을 각 함수 소스에 대어 확률을 반환한다. 한 번에 안 들어가면 요청을 나눈다. 모든 적중은 quickfix에 들어가고, 임계값 이상만 virtual text 표시가 붙는다.","highlightBenefitJa":"正規表現ではなく質問で、SQL を文字列連結するといった言語横断の形を探し、既存の quickfix 編集に載せる。","highlightBenefitKo":"정규식이 아니라 질문으로 SQL을 이어 붙이는 같은 언어 간 모양을 찾고, 기존 quickfix 편집 흐름에 올린다.","evidence":[{"url":"https://github.com/valentynkit/jev.nvim/blob/a1577b0ec874208377e5f38496577d2b8797ab11/README.md","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/valentynkit/jev.nvim/blob/a1577b0ec874208377e5f38496577d2b8797ab11/plugin/jev.lua","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/valentynkit/jev.nvim/blob/a1577b0ec874208377e5f38496577d2b8797ab11/lua/jev/client.lua","note":"固定版本的 Jev 集成与说明来源"}]},{"id":"valentynkit:jev-skip","name":"jev-skip","author":"valentynkit","url":"https://github.com/valentynkit/jev-skip","category":"Decision Tools","plainSummary":"Chrome 扩展：只读字幕，把片段交给 Jev 分类（段数或 Token 预算超限就拆请求），在进度条上画出五类片段，并对达到阈值的 sponsor、self_promo、intro、outro、recap 自动跳过。","plainSummaryEn":"A browser extension that classifies YouTube caption segments with Jev, splitting requests when the segment or token budget is exceeded, paints five categories on the seek bar, and auto-skips those painted slices over the threshold.","jevDecisionPoint":"每个字幕片段一个 Choice：content、sponsor、intro、outro、self_promo、recap 或 other。绘制并跳过的是 PAINTED 五类（sponsor、self_promo、intro、outro、recap）里概率达到阈值的片段；content 和 other 不跳过。","jevDecisionPointEn":"One Choice per caption segment among content, sponsor, intro, outro, self_promo, recap, and other. Painted categories sponsor, self_promo, intro, outro, and recap skip when probability is at or above the threshold; content and other never skip.","highlightBenefit":"不必等 SponsorBlock 人工打点；没有字幕就不判断、也不跳过。","highlightBenefitEn":"Works on videos nobody has labeled yet; with no captions it does nothing.","tags":["typed-decisions","classification-ranking"],"stars":3,"forks":0,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-19T09:38:56Z","createdAt":"2026-09-18T20:16:28Z","metadataFetchedAt":"2026-09-21T22:06:24.009Z","avatarUrl":"https://avatars.githubusercontent.com/u/35431524?v=4","summarySource":"curated","claimStatus":"优先保留投稿者与仓库原文，缺失语言自动补充。自动检查已确认 Jev 集成证据。","claimStatusEn":"Author and repository text is preserved where clear; missing languages are enriched automatically. Checks establish source-level Jev integration, not runtime, safety, or performance validation.","plainSummaryJa":"Chrome 拡張。字幕だけを読み、区間を Jev に分類させる（区間数や Token 予算を超えたらリクエストを分割）。シークバーに五類を描き、しきい値以上の sponsor、self_promo、intro、outro、recap を自動で飛ばす。","plainSummaryKo":"Chrome 확장. 자막만 읽고 구간을 Jev에 분류하며, 구간 수나 Token 예산이 넘으면 요청을 나눈다. 시크바에 다섯 종류를 그리고 임계값 이상인 sponsor, self_promo, intro, outro, recap을 자동으로 건너뛴다.","jevDecisionPointJa":"字幕区間ごとに Choice。content、sponsor、intro、outro、self_promo、recap、other。描画して飛ばすのは PAINTED の五類（sponsor、self_promo、intro、outro、recap）でしきい値以上のもの。content と other は飛ばさない。","jevDecisionPointKo":"자막 구간마다 Choice 하나. content, sponsor, intro, outro, self_promo, recap, other. 그리고 건너뛰는 것은 PAINTED 다섯 종류(sponsor, self_promo, intro, outro, recap) 중 임계값 이상인 구간이다. content와 other는 건너뛰지 않는다.","highlightBenefitJa":"SponsorBlock の人手打点を待たない。字幕がなければ判断もスキップもしない。","highlightBenefitKo":"SponsorBlock 사람 표시를 기다리지 않는다. 자막이 없으면 판단도 건너뛰기도 하지 않는다.","evidence":[{"url":"https://github.com/valentynkit/jev-skip/blob/6837e3e0f1a48cbfc48c85415d99bcfe3eaf0628/README.md","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/valentynkit/jev-skip/blob/6837e3e0f1a48cbfc48c85415d99bcfe3eaf0628/lib/jev.ts","note":"固定版本的 Jev 集成与说明来源"}]},{"id":"valentynkit:jev-plays-pokemon-red","name":"jev-plays-pokemon-red","author":"valentynkit","url":"https://github.com/valentynkit/jev-plays-pokemon-red","category":"Decision Tools","plainSummary":"在 PyBoy 上玩 Pokemon Red：代码负责路线和算术，只在游戏真正分叉时让 Jev 从已合法的动作里选一个。","plainSummaryEn":"Pokemon Red on PyBoy where Python owns the route and the arithmetic, and Jev chooses only among legal actions at actual branches.","jevDecisionPoint":"在合法动作上做 Choice；战斗回合另问本回合是否会倒下、是否该逃跑；无法识别的答案走代码默认。","jevDecisionPointEn":"A Choice over legal actions at each branch; battle turns also ask Noul faints_this_turn and should_flee; failures take the code default.","highlightBenefit":"模型只做闭集选择，不规划整局；失败时不会比脚本更松。","highlightBenefitEn":"The model only picks among code-proven legal moves; a failed call never loosens the script.","tags":["typed-decisions","classification-ranking"],"stars":4,"forks":1,"openIssues":1,"license":"MIT","lastCommitAt":"2026-09-19T09:30:51Z","createdAt":"2026-09-18T20:11:29Z","metadataFetchedAt":"2026-09-21T22:06:27.029Z","avatarUrl":"https://avatars.githubusercontent.com/u/35431524?v=4","summarySource":"curated","claimStatus":"优先保留投稿者与仓库原文，缺失语言自动补充。自动检查已确认 Jev 集成证据。","claimStatusEn":"Author and repository text is preserved where clear; missing languages are enriched automatically. Checks establish source-level Jev integration, not runtime, safety, or performance validation.","plainSummaryJa":"PyBoy 上の Pokemon Red。経路と算術はコードが持ち、ゲームが本当に分岐するときだけ Jev が既に合法な動作から 1 つ選ぶ。","plainSummaryKo":"PyBoy에서 Pokemon Red. 경로와 산술은 코드가 맡고, 게임이 실제로 갈라질 때만 Jev가 이미 적법한 동작 중 하나를 고른다.","jevDecisionPointJa":"合法動作の上で Choice。戦闘ターンは今ターンで倒れるか、逃げるべきかの Noul も問う。認識できない答えはコードの既定に従う。","jevDecisionPointKo":"적법한 동작에서 Choice. 전투 턴에는 이번 턴에 쓰러지는지, 도망가야 하는지도 Noul로 묻는다. 알아볼 수 없는 답은 코드 기본값을 쓴다.","highlightBenefitJa":"モデルは閉集合の選択だけし、局全体は計画しない。失敗してもスクリプトより緩くならない。","highlightBenefitKo":"모델은 닫힌 집합 선택만 하고 판 전체를 계획하지 않는다. 실패해도 스크립트보다 느슨해지지 않는다.","evidence":[{"url":"https://github.com/valentynkit/jev-plays-pokemon-red/blob/cbe5387aeb6c7b3ef6b1f67d4a95e284aee3b0af/README.md","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/valentynkit/jev-plays-pokemon-red/blob/cbe5387aeb6c7b3ef6b1f67d4a95e284aee3b0af/src/jpp/policy.py","note":"固定版本的 Jev 集成与说明来源"}]},{"id":"bugkiwi:elons-job","name":"elons-job","author":"bugkiwi","url":"https://github.com/bugkiwi/elons-job","category":"Context GC & Filter","plainSummary":"本地优先的 Chrome 扩展：先用规则筛选 X 回复，再让 Jev 判断色情、性暗示和引流内容，并提供可恢复的隐藏占位符。","plainSummaryEn":"Local-first Chrome extension that uses Jev to filter sexual and solicitation content in X replies with reversible hidden placeholders.","jevDecisionPoint":"将 X 详情页回复编译为 Jev Noul 判断问题，由本地阈值、规则和重复模板信号共同决定是否隐藏。","jevDecisionPointEn":"Scores X reply text with Jev Noul questions for sexual content, solicitation, and spam, then combines probabilities with local thresholds and structural signals.","highlightBenefit":"结合本地规则、缓存、并发与成本保护以及 Fail-Open 设计；不需要 X 官方 API，隐藏内容可以恢复。","highlightBenefitEn":"Combines local rules, caching, concurrency and cost controls, and fail-open handling; it requires no X API and keeps hidden comments recoverable.","tags":["context-compaction","browser-automation","security-guardrails"],"stars":10,"forks":0,"openIssues":0,"license":"MIT","language":"JavaScript","lastCommitAt":"2026-09-20T00:47:04Z","createdAt":"2026-09-18T05:46:42Z","metadataFetchedAt":"2026-09-21T22:06:30.069Z","avatarUrl":"https://avatars.githubusercontent.com/u/1936250?v=4","summarySource":"source-reviewed","claimStatus":"README 与源码展示 TypeSafe System One 接入及确定性 mock 测试。","claimStatusEn":"The README and source show TypeSafe System One integration and deterministic mock tests; not independently run or performance-tested by this site.","evidence":[{"url":"https://github.com/bugkiwi/elons-job/blob/be01f869871fcda5841fb77df3df6e46abbda493/README.md","note":"固定版本的 Jev 集成与项目说明"},{"url":"https://github.com/bugkiwi/elons-job/blob/be01f869871fcda5841fb77df3df6e46abbda493/src/background.js#L46-L101","note":"TypeSafe System One 客户端与 Noul 请求实现"},{"url":"https://github.com/bugkiwi/elons-job/blob/be01f869871fcda5841fb77df3df6e46abbda493/src/shared/core.js#L220-L327","note":"Jev Noul 规则编译、概率解析与阈值判定"}],"pinned":false},{"id":"bugkiwi:turing-jail","name":"turing-jail","author":"bugkiwi","url":"https://github.com/bugkiwi/turing-jail","category":"Decision Tools","plainSummary":"由 TypeSafe Jev System One 驱动的三关 AI 审讯游戏：通过求情、逻辑与悖论测试，争取获得释放。","plainSummaryEn":"Interactive three-level AI interrogation game powered by TypeSafe Jev; write responses and pass plea, logic, and paradox verdicts to earn release.","jevDecisionPoint":"每一关评估 prisoner_response 的释放概率、求情、逻辑与悖论信号，同时用 Choice 识别说服策略、用 Score 评估说服力。","jevDecisionPointEn":"For each level, Jev evaluates release probability plus plea, logic, and paradox signals, then selects a persuasion tactic and scores persuasiveness.","highlightBenefit":"把结构化 Jev 判断变成可玩的反馈、门槛和排行榜结果，让用户直观看见不同论证如何影响释放概率。","highlightBenefitEn":"Turns structured Jev judgments into playable feedback, pass thresholds, and leaderboard results that show how arguments affect release probability.","tags":["typed-decisions","games-simulation"],"stars":0,"forks":0,"openIssues":0,"license":null,"language":"TypeScript","lastCommitAt":"2026-09-18T03:18:57Z","createdAt":"2026-09-17T08:23:53Z","metadataFetchedAt":"2026-09-21T22:06:33.064Z","avatarUrl":"https://avatars.githubusercontent.com/u/1936250?v=4","summarySource":"human-reviewed","claimStatus":"开源 TypeScript Web 游戏，支持中文、英文和德文，并将每次完成的三关结果保存为可分享记录。","claimStatusEn":"Source-reviewed integration record; runtime behavior and performance were not independently retested by this site.","evidence":[{"url":"https://github.com/bugkiwi/turing-jail/blob/629ff1b5617800d851811ce4379845f5ac59f729/README.md","note":"固定版本的项目说明"},{"url":"https://github.com/bugkiwi/turing-jail/blob/629ff1b5617800d851811ce4379845f5ac59f729/server.ts#L91-L123","note":"固定版本的 Jev 集成与源码实现"}],"pinned":false},{"id":"typesafe-ai:typesafe-sdk-python","name":"typesafe-sdk-python","author":"typesafe-ai","url":"https://github.com/typesafe-ai/typesafe-sdk-python","category":"SDK & Decision Frameworks","plainSummary":"TypeSafe 官方 Python SDK，提供 Jev System One 的同步、异步客户端及问题和回答类型。","plainSummaryEn":"Official TypeSafe Python SDK with synchronous and asynchronous clients for Jev System One, plus question and answer types.","jevDecisionPoint":"system_one 将上下文与选择、评分或是非问题提交到 API，并按问题名称解析返回的回答。","jevDecisionPointEn":"system_one sends context and choice, score or yes/no questions to the API, then parses answers by question name.","highlightBenefit":"在 Python 中复用请求、类型化回答和连接管理，支持 with 与 async with。","highlightBenefitEn":"Reuses request handling, typed answers and connection management in Python, with support for with and async with.","tags":["multilanguage-sdk","typed-decisions"],"stars":184,"forks":23,"openIssues":2,"license":"MIT","language":"Python","lastCommitAt":"2026-09-21T15:57:06Z","createdAt":"2026-09-04T21:01:31Z","metadataFetchedAt":"2026-09-21T22:06:36.043Z","avatarUrl":"https://avatars.githubusercontent.com/u/171090879?v=4","summarySource":"human-reviewed","claimStatus":"已核对固定版本 README、客户端源码与 MIT 许可证。","claimStatusEn":"Reviewed the fixed-version README, client source and MIT license.","plainSummaryJa":"TypeSafe 公式の Python SDK。Jev System One の同期・非同期クライアントと、質問・回答の型を提供する。","plainSummaryKo":"TypeSafe 공식 Python SDK로, Jev System One의 동기 및 비동기 클라이언트와 질문·응답 타입을 제공한다.","jevDecisionPointJa":"system_one が文脈と選択・採点・はい／いいえの質問を API に送り、質問名ごとに回答を読み取る。","jevDecisionPointKo":"system_one이 맥락과 선택·점수·예/아니요 질문을 API에 보내고, 질문 이름별로 응답을 파싱한다.","highlightBenefitJa":"Python でリクエスト処理、型付き回答、接続管理を共用でき、with と async with に対応する。","highlightBenefitKo":"Python에서 요청 처리, 타입이 있는 응답, 연결 관리를 재사용하며 with와 async with를 지원한다.","claimStatusJa":"固定バージョンの README、クライアント実装、MIT ライセンスを確認済み。実行および性能・安全性の独立検証は行っていない。","claimStatusKo":"고정 버전의 README, 클라이언트 소스와 MIT 라이선스를 확인했다. 직접 실행하거나 성능 및 보안을 독립적으로 검증하지는 않았다.","licenseStatus":"confirmed","sourceReviewedAt":"2026-09-19T10:34:33.417699+00:00","evidence":[{"url":"https://github.com/typesafe-ai/typesafe-sdk-python/blob/2ce5c65f13646cab6e6f782328194c9d85f3300a/src/typesafe_sdk/_core/client/sync/client.py#L206-L221","note":"同步客户端提交状态与问题，并解析类型化响应 / Synchronous client submits state and questions and parses typed responses"},{"url":"https://github.com/typesafe-ai/typesafe-sdk-python/blob/2ce5c65f13646cab6e6f782328194c9d85f3300a/src/typesafe_sdk/_core/client/aio/client.py#L216-L243","note":"异步客户端及连接生命周期 / Asynchronous client and connection lifecycle"},{"url":"https://github.com/typesafe-ai/typesafe-sdk-python/blob/2ce5c65f13646cab6e6f782328194c9d85f3300a/src/typesafe_sdk/_core/question_types.py#L77-L113","note":"选择、评分与是非问题类型 / Choice, score and yes/no question types"},{"url":"https://github.com/typesafe-ai/typesafe-sdk-python/blob/2ce5c65f13646cab6e6f782328194c9d85f3300a/src/typesafe_sdk/constants.py#L18-L19","note":"默认模型为 jev-latest / Default model is jev-latest"},{"url":"https://github.com/typesafe-ai/typesafe-sdk-python/blob/2ce5c65f13646cab6e6f782328194c9d85f3300a/README.md#L13-L29","note":"固定版本快速入门 / Fixed-version quickstart"},{"url":"https://github.com/typesafe-ai/typesafe-sdk-python/blob/2ce5c65f13646cab6e6f782328194c9d85f3300a/LICENSE#L1-L21","note":"固定版本 MIT 许可证全文 / Fixed-version MIT license text"}],"pinned":false},{"id":"grayrepo-byte:jev_filter_for_x","name":"JevFilterForX","author":"grayrepo-byte","url":"https://github.com/grayrepo-byte/jev_filter_for_x","category":"Browser & OS Action","plainSummary":"JevFilterForX 是一个用于 X 的浏览器扩展。它会为信息流中的帖子评分，用轻量标签解释评分，并根据你的过滤设置折叠内容。未配置 API Key 时，默认使用本地模拟评分。","plainSummaryEn":"A browser extension that scores and filters X posts in real time with Jev, folding low-signal content while keeping it expandable. Without an API key, it defaults to local mock scoring.","jevDecisionPoint":"Jev 用 Choice 分类、Score 评估信息量、可行动性与原创性，用 Noul 打标签；本地阈值和噪声规则决定是否折叠帖子。","jevDecisionPointEn":"Jev classifies posts with Choice, rates signal, actionability and originality with Score, and assigns Noul labels; local thresholds and noise rules decide whether to fold each post.","highlightBenefit":"把评分、标签和可调整的过滤阈值放进 X 信息流，折叠后仍可展开或再次隐藏帖子及其媒体附件。","highlightBenefitEn":"Adds scores, labels and adjustable filtering thresholds to the X feed, with reversible folding of posts and their attached media.","tags":["browser-automation","classification-ranking"],"stars":1,"forks":0,"openIssues":0,"license":null,"language":"TypeScript","lastCommitAt":"2026-09-20T13:47:21Z","createdAt":"2026-09-19T07:27:19Z","metadataFetchedAt":"2026-09-21T22:06:39.069Z","avatarUrl":"https://avatars.githubusercontent.com/u/179504272?v=4","summarySource":"human-reviewed","claimStatus":"仅核对固定提交源码，未运行或验证准确率、延迟、成本。无 Key 的模拟结果与真实结果共用 30 天缓存，补 Key 后不会重算既有缓存；兴趣/价值偏好未接入过滤，缺失的 API 评分按 0 分处理。联网模式向 TypeSafe 发送帖子文本、作者和 ID；Key 存本机。公开源码未发现许可证。","claimStatusEn":"Fixed-commit source review only; no runtime, accuracy, latency or cost validation. Mock and live results share a 30-day cache that is not invalidated when a key is added. Interest/value preferences do not affect filtering; missing API scores become zero. Online mode sends post text, author and ID to TypeSafe; the key is stored locally. Public source has no identified license.","plainSummaryJa":"X の投稿を Jev で採点し、ラベルを表示して、フィルターに合う投稿を再展開可能な形で折りたたむブラウザー拡張。API キー未設定時はローカルの模擬採点が既定で使われる。","plainSummaryKo":"Jev로 X 게시물에 점수를 매기고 라벨을 표시하며 필터에 해당하는 게시물을 다시 펼칠 수 있게 접는 브라우저 확장. API 키가 없으면 기본적으로 로컬 모의 점수를 사용한다.","jevDecisionPointJa":"Jev の Choice で投稿を分類し、Score で情報量・実行可能性・独自性を採点、Noul でラベルを付ける。ローカルの閾値とノイズ規則で折りたたみを決める。","jevDecisionPointKo":"Jev의 Choice로 게시물을 분류하고 Score로 정보량, 실행 가능성과 독창성을 평가하며 Noul로 라벨을 붙인다. 로컬 임계값과 노이즈 규칙으로 게시물 접기를 결정한다.","highlightBenefitJa":"X のタイムラインに採点、ラベル、調整可能な閾値を加え、投稿と添付メディアを折りたたんだ後も再表示・再非表示にできる。","highlightBenefitKo":"X 피드에 점수, 라벨, 조정 가능한 필터 임계값을 추가하고 게시물과 첨부 미디어를 접은 뒤 다시 펼치거나 숨길 수 있다.","claimStatusJa":"固定コミットのソース確認のみで、実行・精度・遅延・費用は未検証。模擬結果と API 結果は同じ 30 日キャッシュに入り、キー追加時も既存結果を再採点しない。興味・価値の設定はフィルターに未接続で、API の採点欠落は 0 点になる。オンライン時は投稿本文・投稿者・ID を TypeSafe に送り、キーは端末内に保存する。公開ソースにライセンスは見つからない。","claimStatusKo":"고정 커밋 소스만 검토했으며 실행, 정확도, 지연 시간, 비용은 검증하지 않았다. 모의 결과와 실제 결과가 같은 30일 캐시를 쓰며 키를 추가해도 기존 결과를 다시 평가하지 않는다. 관심사와 가치 선호는 필터에 반영되지 않고 누락된 API 점수는 0점이 된다. 온라인 모드에서는 게시물 본문, 작성자, ID를 TypeSafe에 보내며 키는 기기에 저장한다. 공개 소스에서 라이선스를 찾지 못했다.","licenseStatus":"unconfirmed","sourceReviewedAt":"2026-09-19T12:23:06.263963+00:00","evidence":[{"url":"https://github.com/grayrepo-byte/jev_filter_for_x/blob/62cb4a602026b5449a9caa94c193c534e0a19c01/src/core/jev.ts#L35-L255","note":"Reachable direct Jev HTTP request, typed questions and answer mapping; source review only."},{"url":"https://github.com/grayrepo-byte/jev_filter_for_x/blob/62cb4a602026b5449a9caa94c193c534e0a19c01/src/core/classifier.ts#L28-L66","note":"No API key selects a deterministic text-hash mock; a key selects JevClassifier."},{"url":"https://github.com/grayrepo-byte/jev_filter_for_x/blob/62cb4a602026b5449a9caa94c193c534e0a19c01/src/core/policy.ts#L11-L47","note":"Local visibility branches: focus score, selected noise categories and hideNoise/score."},{"url":"https://github.com/grayrepo-byte/jev_filter_for_x/blob/62cb4a602026b5449a9caa94c193c534e0a19c01/src/entrypoints/background/index.ts#L21-L43","note":"Background handler looks up cached classifications before choosing mock or Jev."},{"url":"https://github.com/grayrepo-byte/jev_filter_for_x/blob/62cb4a602026b5449a9caa94c193c534e0a19c01/src/storage/cache.ts#L7-L64","note":"30-day tweet-ID cache does not distinguish mock and live results."},{"url":"https://github.com/grayrepo-byte/jev_filter_for_x/blob/62cb4a602026b5449a9caa94c193c534e0a19c01/src/storage/settings.ts#L4-L44","note":"Enabled by default; API key is device-local and setting it does not invalidate cache."},{"url":"https://github.com/grayrepo-byte/jev_filter_for_x/blob/62cb4a602026b5449a9caa94c193c534e0a19c01/src/ui/options/SignalFeedOptions.tsx#L117-L165","note":"Interest/value controls exist, but those settings are not consumed by visibility policy."}],"pinned":false},{"id":"hqman:jevscout","name":"JevScout","author":"hqman","url":"https://github.com/hqman/JevScout","category":"Domain & Vertical Tools","plainSummary":"供编程 Agent 调用的求职检索演示 Skill，通过 Chrome 浏览公司招聘页，用 Jev 筛选 AI 与软件工程职位并保存结果。","plainSummaryEn":"A demo job-search Skill for coding Agents that browses company careers pages in Chrome, uses Jev to screen AI and software-engineering roles, and saves the results.","jevDecisionPoint":"Jev 用 Choice 识别页面类型，用 Noul 评估招聘链接、职位相关性、筛选控件和候选人匹配度；本地阈值决定导航、打开职位与保存。","jevDecisionPointEn":"Jev uses Choice to classify pages and Noul to assess careers links, role relevance, filter controls and candidate fit; local thresholds select navigation, job details and saved results.","highlightBenefit":"把招聘入口、职位筛选和详情匹配连成一次 CLI 流程，输出本地 JSON 与 Markdown 报告。","highlightBenefitEn":"Connects careers navigation, role screening and detail matching in one CLI workflow, with local JSON and Markdown reports.","tags":["browser-automation","domain-workflows","classification-ranking"],"stars":28,"forks":6,"openIssues":1,"license":null,"language":"Python","lastCommitAt":"2026-09-18T13:14:21Z","createdAt":"2026-09-18T13:07:05Z","metadataFetchedAt":"2026-09-21T22:06:42.019Z","avatarUrl":"https://avatars.githubusercontent.com/u/185824?v=4","summarySource":"human-reviewed","claimStatus":"演示版 MVP，仅核对固定提交源码，未运行浏览器或真实 API。扫描有数量与文本上限；mock 的分数和延迟为模拟值，仍可操作真实 Chrome。不能视为严格只读流程；跨站覆盖、匹配准确率与性能未验证，未确认开源许可。","claimStatusEn":"Demo MVP reviewed at a fixed commit; no browser or live API run. Scanning has count and text limits. Mock scores and latency are simulated, while mock mode can still operate real Chrome. The workflow is not guaranteed read-only. Site coverage, matching accuracy and performance were not validated; no open-source license was confirmed.","plainSummaryJa":"Chrome で企業の採用ページを巡り、Jev で AI・ソフトウェア開発職を選別して結果を保存する、コーディング Agent 向けの求人検索デモ Skill。","plainSummaryKo":"Chrome으로 기업 채용 페이지를 탐색하고 Jev로 AI 및 소프트웨어 개발 직무를 선별해 결과를 저장하는 코딩 Agent용 구직 검색 데모 Skill.","jevDecisionPointJa":"Jev の Choice でページ種別を判定し、Noul で採用リンク、職種の関連性、絞り込み操作、候補者との適合度を評価。ローカルの閾値で移動、詳細表示、保存を決める。","jevDecisionPointKo":"Jev의 Choice로 페이지 유형을 분류하고 Noul로 채용 링크, 직무 관련성, 필터 요소와 지원자 적합도를 평가한다. 로컬 임계값으로 이동, 상세 조회와 저장을 결정한다.","highlightBenefitJa":"採用ページへの移動、職種の選別、詳細の照合を一つの CLI フローにつなぎ、JSON と Markdown のレポートをローカルに出力する。","highlightBenefitKo":"채용 페이지 탐색, 직무 선별과 상세 적합도 평가를 하나의 CLI 흐름으로 연결하고 JSON과 Markdown 보고서를 로컬에 저장한다.","claimStatusJa":"固定コミットのソースのみ確認したデモ MVP。ブラウザーや実 API は未実行。走査件数と文字数に上限があり、mock のスコアと遅延は模擬値だが、実際の Chrome は操作し得る。読み取り専用の動作は保証されない。サイト対応範囲、適合精度、性能は未検証で、オープンソースライセンスも未確認。","claimStatusKo":"고정 커밋의 소스만 검토한 데모 MVP이며 브라우저나 실제 API는 실행하지 않았다. 탐색 개수와 텍스트 길이에 제한이 있다. mock 점수와 지연 시간은 모의 값이지만 실제 Chrome은 조작할 수 있다. 읽기 전용 동작은 보장되지 않는다. 사이트 지원 범위, 적합도 정확성과 성능은 검증하지 않았으며 오픈소스 라이선스도 확인되지 않았다.","licenseStatus":"unconfirmed","sourceReviewedAt":"2026-09-19T12:22:02.798187+00:00","evidence":[{"url":"https://github.com/hqman/JevScout/blob/b51390be0c3ca28b5407ab6d8c8044912a03bb31/README.md#L1-L30","note":"README.md"},{"url":"https://github.com/hqman/JevScout/blob/b51390be0c3ca28b5407ab6d8c8044912a03bb31/jev_job_hunter/jev.py#L220-L305","note":"jev_job_hunter/jev.py"},{"url":"https://github.com/hqman/JevScout/blob/b51390be0c3ca28b5407ab6d8c8044912a03bb31/jev_job_hunter/questions.py#L3-L88","note":"jev_job_hunter/questions.py"},{"url":"https://github.com/hqman/JevScout/blob/b51390be0c3ca28b5407ab6d8c8044912a03bb31/jev_job_hunter/runner.py#L70-L112","note":"jev_job_hunter/runner.py"},{"url":"https://github.com/hqman/JevScout/blob/b51390be0c3ca28b5407ab6d8c8044912a03bb31/jev_job_hunter/policy.py#L88-L177","note":"jev_job_hunter/policy.py"}],"pinned":false},{"id":"saibimajdi:typesafeai-dotnet-sdk","name":"typesafeai-dotnet-sdk","author":"saibimajdi","url":"https://github.com/saibimajdi/typesafeai-dotnet-sdk","category":"SDK & Decision Frameworks","plainSummary":"适用于 .NET 8+ 的 TypeSafe AI / Jev 客户端 SDK，支持 Choice、Score 与 Noul 决策原语与强类型响应解析。","plainSummaryEn":"Community .NET SDK for TypeSafe AI and Jev, providing asynchronous typed evaluation for Choice, Score, and Noul primitives.","jevDecisionPoint":"在 TypeSafeClient 中封装 System One API 请求，以并行 HTTP 管道评估 typed questions 并映射为不可变 C# 记录。","jevDecisionPointEn":"Wraps System One API endpoints in TypeSafeClient to issue parallel question evaluations and parse responses into immutable C# records.","highlightBenefit":"为 C# / .NET 生态提供原生的 Jev 异步客户端支持，具备依赖注入扩展、指数退避重试与强类型错误处理。","highlightBenefitEn":"Brings native Jev decision primitives to the .NET ecosystem with dependency injection integration and configurable retry policies.","tags":["multilanguage-sdk","typed-decisions"],"stars":6,"forks":0,"openIssues":2,"license":"MIT","language":"C#","lastCommitAt":"2026-09-18T19:18:07Z","createdAt":"2026-09-16T19:18:00Z","metadataFetchedAt":"2026-09-21T22:06:45.029Z","avatarUrl":"https://avatars.githubusercontent.com/u/8198978?v=4","summarySource":"human-reviewed","claimStatus":"源码已核验。包含完整的客户端实现与类型定义；未与远程生产服务联调，不包含官方商业保证。","claimStatusEn":"Source code verified against pinned commit. Complete client implementation and models; no live production service benchmark.","plainSummaryJa":"TypeSafe AI および Jev 向けの .NET 向けコミュニティ SDK であり、Choice、Score、Noul などの型付き判断を非同期で評価します。","plainSummaryKo":"TypeSafe AI 및 Jev를 위한 .NET 커뮤니티 SDK로 Choice, Score, Noul 판단 기본 요소를 비동기적으로 평가합니다.","jevDecisionPointJa":"TypeSafeClient 内で System One API エンドポイントをラップし、並列質問評価と型付きレスポンスのパースを行います。","jevDecisionPointKo":"TypeSafeClient에서 System One API 호출을 캡슐화하여 병렬 질문 평가 및 불변 C# 레코드 매핑을 수행합니다.","highlightBenefitJa":"依存性注入や指数バックオフ再試行を備えた、.NET 向けネイティブな Jev 統合クライアントを提供します。","highlightBenefitKo":"의존성 주입 지원과 지수 백오프 재시도를 갖춘 .NET 네이티브 Jev 클라이언트를 제공합니다.","claimStatusJa":"固定コミットのソースを確認済み。クライアント実装とモデル定義を含むが、本番環境での実稼働ベンチマークは未実施。","claimStatusKo":"고정 커밋 기준 소스 확인 완료. 클라이언트 구현 및 모델을 포함하나 상용 서비스 연동 평가는 수행하지 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T13:00:00.000000+00:00","evidence":[{"url":"https://github.com/saibimajdi/typesafeai-dotnet-sdk/blob/6be669e0d0b116b76e85136c4f6eaa44fadcecab/src/TypeSafe.Sdk/TypeSafeClient.cs#L1-L283","note":"src/TypeSafe.Sdk/TypeSafeClient.cs (L1-L283)"},{"url":"https://github.com/saibimajdi/typesafeai-dotnet-sdk/blob/6be669e0d0b116b76e85136c4f6eaa44fadcecab/README.md","note":"README.md"}],"pinned":false,"topics":["ai","classification","csharp","dotnet","llm","sdk","systemone","typesafe"]},{"id":"jamesward:zio-typesafe-ai","name":"zio-typesafe-ai","author":"jamesward","url":"https://github.com/jamesward/zio-typesafe-ai","category":"SDK & Decision Frameworks","plainSummary":"Scala 3 / ZIO 的 Jev 客户端：用 NamedTuple 一次提交多个 Noul、Choice、Score，答案按同样字段名返回。","plainSummaryEn":"A Scala 3 / ZIO client that asks typed Noul, Choice, and Score questions in one round-trip and returns a NamedTuple of answers.","jevDecisionPoint":"把状态和类型化问题编成 System One 请求，解码 Probability、Choice 分布和 Score。","jevDecisionPointEn":"Encodes state and typed questions as a System One request and decodes Probability, Choice distributions, and Score values.","highlightBenefit":"在 ZIO 里用编译期字段名对接 Jev，避免手写 JSON 和字符串键查找。","highlightBenefitEn":"Keeps Jev questions and answers as compile-time field names in ZIO, without raw JSON key lookups.","tags":["multilanguage-sdk","typed-decisions"],"stars":3,"forks":0,"openIssues":0,"license":"Apache-2.0","language":"Scala","lastCommitAt":"2026-09-18T14:32:10Z","createdAt":"2026-09-17T05:23:30Z","metadataFetchedAt":"2026-09-21T22:06:48.028Z","avatarUrl":"https://avatars.githubusercontent.com/u/65043?v=4","summarySource":"curated","claimStatus":"源码已核验。包含完整的 ZIO 封装与测试规格；未执行生产级负载基准测试。","claimStatusEn":"Source code verified against pinned commit. Complete effect wrapper and test specifications; no live load benchmark.","plainSummaryJa":"Scala 3 / ZIO の Jev クライアント。NamedTuple で複数の Noul、Choice、Score を一度に送り、同じフィールド名で答えを返す。","plainSummaryKo":"Scala 3 / ZIO Jev 클라이언트. NamedTuple로 Noul, Choice, Score를 한 번에 보내고 같은 필드 이름으로 답을 받는다.","jevDecisionPointJa":"状態と型付き質問を System One リクエストに組み立て、Probability、Choice 分布、Score をデコードする。","jevDecisionPointKo":"상태와 타입이 있는 질문을 System One 요청으로 만들어 Probability, Choice 분포, Score를 디코딩한다.","highlightBenefitJa":"ZIO 上でコンパイル時のフィールド名として Jev に接続し、手書き JSON や文字列キー参照を避ける。","highlightBenefitKo":"ZIO에서 컴파일 타임 필드 이름으로 Jev에 연결해 손으로 쓴 JSON과 문자열 키 조회를 피한다.","claimStatusJa":"固定コミットのソースを確認済み。ZIO ラッパーとテスト仕様を含むが、高負荷ベンチマークは未実施。","claimStatusKo":"고정 커밋 기준 소스 확인 완료. ZIO 래퍼 및 테스트 규격을 포함하나 고부하 벤치마크는 미실시되었습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T13:00:00.000000+00:00","evidence":[{"url":"https://github.com/jamesward/zio-typesafe-ai/blob/38082e8712b49ce239ae628557619347909165dc/src/main/scala/com/jamesward/zio_typesafe_ai/TypeSafeAI.scala#L1-L571","note":"src/main/scala/com/jamesward/zio_typesafe_ai/TypeSafeAI.scala (L1-L571)"},{"url":"https://github.com/jamesward/zio-typesafe-ai/blob/38082e8712b49ce239ae628557619347909165dc/README.md","note":"README.md"}],"pinned":false},{"id":"genierobot:typesafe-ai-rails","name":"typesafe-ai-rails","author":"GenieRobot","url":"https://github.com/GenieRobot/typesafe-ai-rails","category":"SDK & Decision Frameworks","plainSummary":"Ruby on Rails 官方风格集成插件，为 ActiveModel/ActiveRecord 模型引入 Jev 分类、评分与决策策略支持。","plainSummaryEn":"Ruby on Rails integration gem for TypeSafe AI and Jev, providing model-level classification and decision policy patterns.","jevDecisionPoint":"在 Rails 模型生命周期或服务层中调用 Jev System One API，执行业务状态判断并自动记录决策日志。","jevDecisionPointEn":"Dispatches state and criteria from ActiveRecord models or services to Jev System One endpoints, logging structured outcomes.","highlightBenefit":"极简化 Rails 应用接入 Jev 的流程，提供 Railtie 自动加载、配置生成器与结构化决策策略模式。","highlightBenefitEn":"Simplifies adding Jev discrete judgments to Rails web apps with standard Railtie configuration and policy wrappers.","tags":["multilanguage-sdk","domain-workflows","typed-decisions"],"stars":2,"forks":0,"openIssues":0,"license":"MIT","language":"Ruby","lastCommitAt":"2026-09-16T20:55:50Z","createdAt":"2026-09-16T20:31:02Z","metadataFetchedAt":"2026-09-21T22:06:51.011Z","avatarUrl":"https://avatars.githubusercontent.com/u/19553316?v=4","summarySource":"human-reviewed","claimStatus":"源码已核验。包含 Rails 集成脚手架、客户端封装与策略定义；未在外部生产数据库进行集成基准测试。","claimStatusEn":"Source code verified against pinned commit. Includes Railtie setup and policy classes; production database performance unverified.","plainSummaryJa":"Ruby on Rails 向けの TypeSafe AI / Jev 統合 gem であり、モデル層やサービス層に型付き分類と決定ポリシーパターンを導入します。","plainSummaryKo":"TypeSafe AI 및 Jev를 위한 Ruby on Rails 연동 젬으로 모델 계층에 분류 및 판단 정책 패턴을 제공합니다.","jevDecisionPointJa":"ActiveRecord モデルから状態と判定基準を Jev System One エンドポイントへ送り、構造化された判断ログを記録します。","jevDecisionPointKo":"ActiveRecord 모델에서 상태와 기준을 Jev System One 엔드포인트로 전달하고 구조화된 판단 결과를 기록합니다.","highlightBenefitJa":"標準的な Railtie 設定とポリシーラッパーにより、Rails アプリケーションへの Jev 導入を簡素化します。","highlightBenefitKo":"표준 Railtie 설정과 정책 래퍼를 통해 Rails 웹 애플리케이션에서 간편하게 Jev 판단을 활용할 수 있습니다.","claimStatusJa":"固定コミットのソースを確認済み。Railtie 設定とポリシークラスを含むが、本番データベースでの検証は未実施。","claimStatusKo":"고정 커밋 기준 소스 확인 완료. Railtie 설정 및 정책 클래스를 포함하나 실제 프로덕션 DB 부하 테스트는 미실시되었습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T13:00:00.000000+00:00","evidence":[{"url":"https://github.com/GenieRobot/typesafe-ai-rails/blob/fbac64470904fef7c4e68da172813daf0840f31f/lib/typesafe/rails/client.rb#L1-L148","note":"lib/typesafe/rails/client.rb (L1-L148)"},{"url":"https://github.com/GenieRobot/typesafe-ai-rails/blob/fbac64470904fef7c4e68da172813daf0840f31f/README.md","note":"README.md"}],"pinned":false},{"id":"akashpriyadarshii:jev-superpowers","name":"jev-superpowers","author":"AkashPriyadarshii","url":"https://github.com/AkashPriyadarshii/jev-superpowers","category":"CLI & Pipelines","plainSummary":"面向编码智能体的系统化开发流程框架，融合 Jev 进行无幻觉依赖审查、完成度门禁与错误重试决策。","plainSummaryEn":"Software development framework for coding agents enhanced with Jev for dependency vetting, completion gates, and debugging decisions.","jevDecisionPoint":"在智能体执行循环的关键检查点，利用 Jev 评估代码变更质量、测试覆盖与第三方包合法性，决定是否推进或回滚。","jevDecisionPointEn":"Invokes Jev at workflow checkpoints to evaluate change validity, test sufficiency, and package safety before advancing agent state.","highlightBenefit":"将确定性代码流水线与 Jev 快速二元判断结合，防止编码 Agent 在长程任务中产生方向偏离或引入未知包。","highlightBenefitEn":"Constrains coding agent trajectories using fast discrete judgments, preventing hallucinated dependencies and premature task completion.","tags":["coding-agents","cli-git-gates"],"stars":13,"forks":2,"openIssues":2,"license":"MIT","language":"JavaScript","lastCommitAt":"2026-09-21T17:52:48Z","createdAt":"2026-09-18T12:59:06Z","metadataFetchedAt":"2026-09-21T22:06:54.034Z","avatarUrl":"https://avatars.githubusercontent.com/u/272530059?v=4","summarySource":"human-reviewed","claimStatus":"源码已核验。包含技能脚手架与决策逻辑用例；实际门禁效果取决于接入的模型与具体上下文。","claimStatusEn":"Source code verified against pinned commit. Framework definitions and examples inspected; accuracy depends on upstream model calibration.","plainSummaryJa":"コーディングエージェント向けの体系的開発フレームワークであり、依存関係検証、完了判定ゲート、デバッグ分岐に Jev を活用します。","plainSummaryKo":"의존성 검증, 완료 게이트, 디버깅 분기 제어에 Jev를 결합한 코딩 에이전트용 체계적 개발 프레임워크입니다.","jevDecisionPointJa":"ワークフローのチェックポイントで Jev を呼び出し、変更の妥当性やテスト充足度を判定してエージェントの進行を制御します。","jevDecisionPointKo":"워크플로 검사 지점에서 Jev를 호출하여 변경 사항의 타당성과 테스트 충족도를 검증한 뒤 에이전트 진행을 제어합니다.","highlightBenefitJa":"高速な離散判断によりエージェントの逸脱を防ぎ、未知の依存関係の混入やタスクの誤認完了を抑止します。","highlightBenefitKo":"신속한 이산 판단을 통해 에이전트의 경로 이탈을 막고 잘못된 패키지 설치 및 조기 완료 오류를 차단합니다.","claimStatusJa":"固定コミットのソースを確認済み。フレームワーク定義と使用例を確認。実際の精度は連携モデルの品質に依存します。","claimStatusKo":"고정 커밋 기준 소스 확인 완료. 프레임워크 정의 및 예제를 검토함. 실제 성능은 연동 모델의 상태에 좌우됩니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T13:00:00.000000+00:00","evidence":[{"url":"https://github.com/AkashPriyadarshii/jev-superpowers/blob/6ddc702c231fd33fdf04b1ab1c0cce96fd8d2a19/skills/systematic-debugging/condition-based-waiting-example.ts#L1-L159","note":"skills/systematic-debugging/condition-based-waiting-example.ts (L1-L159)"},{"url":"https://github.com/AkashPriyadarshii/jev-superpowers/blob/6ddc702c231fd33fdf04b1ab1c0cce96fd8d2a19/README.md","note":"README.md"}],"pinned":false,"topics":["agentic-skills","ai-agents","antigravity","claude-code","codex","coding-agents","cursor","developer-tools","gemini-cli","git-jev","jev","jev-scout","software-engineering","superpowers","system-one","tdd","typesafe-ai","zero-hallucination"]},{"id":"morcoan:jevseek","name":"JevSeek","author":"morcoan","url":"https://github.com/morcoan/JevSeek","category":"Domain & Vertical Tools","plainSummary":"本地编码桌面与 CLI 智能体，将 Jev 的动作路由与 DeepSeek 的代码参数生成分层解耦协同。","plainSummaryEn":"Local coding workspace and agent decoupling tool routing via Jev from detailed argument generation via DeepSeek.","jevDecisionPoint":"由 Jev 根据当前工作区状态与用户意图快速判定下一步工具类型，再交由大语言模型补全具体参数。","jevDecisionPointEn":"Employs Jev to classify and select the next tool action based on execution state, delegating parameter generation to DeepSeek.","highlightBenefit":"结合了 Jev 的毫秒级离散路由能力与生成模型的强代码生成能力，降低端到端思考延迟并节省 Token 消耗。","highlightBenefitEn":"Combines low-latency discrete routing with capable generative modeling, reducing overall prompt overhead and turn latency.","tags":["coding-agents","desktop-os"],"stars":3,"forks":0,"openIssues":0,"license":"MIT","language":"Python","lastCommitAt":"2026-09-20T22:24:02Z","createdAt":"2026-09-18T09:06:19Z","metadataFetchedAt":"2026-09-21T22:07:00.070Z","avatarUrl":"https://avatars.githubusercontent.com/u/254227521?v=4","summarySource":"human-reviewed","claimStatus":"源码已核验。包含桌面 UI 桥接与双模型路由逻辑；端到端延迟与实际生成质量依赖本地与远程环境配置。","claimStatusEn":"Source code verified against pinned commit. Covers desktop bridge and routing logic; end-to-end latency depends on provider configuration.","plainSummaryJa":"ローカルコーディングデスクトップ環境およびエージェントであり、Jev によるツールルーティングと DeepSeek によるコード引数生成を分離協調させます。","plainSummaryKo":"로컬 코딩 데스크톱 에이전트로 Jev의 도구 라우팅과 DeepSeek의 인자 생성을 계층화하여 결합합니다.","jevDecisionPointJa":"実行状態と意図に基づき Jev が次のツールアクションを瞬時に選定し、詳細な引数生成のみを DeepSeek に委ねます。","jevDecisionPointKo":"실행 상태와 사용자 의도에 따라 Jev가 다음 도구를 신속하게 선택하고 세부 파라미터 생성만 대규모 언어 모델에 위임합니다.","highlightBenefitJa":"低遅延な離散ルーティングと高性能な生成モデルを組み合わせ、プロンプト消費とターン遅延を削減します。","highlightBenefitKo":"지연 시간이 짧은 이산 라우팅과 강력한 생성 능력을 결합하여 턴당 지연 시간과 토큰 소모를 줄입니다.","claimStatusJa":"固定コミットのソースを確認済み。デスクトップブリッジと連携ロジックを含むが、総合性能は各環境の設定に依存します。","claimStatusKo":"고정 커밋 기준 소스 확인 완료. 데스크톱 연동 및 라우팅 로직을 확인했으나 종합 지연 시간은 환경 설정에 따라 달라집니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T13:00:00.000000+00:00","evidence":[{"url":"https://github.com/morcoan/JevSeek/blob/030d59114a03720783e839481a7814e7aecaad85/desktop.py#L1-L141","note":"desktop.py (L1-L141)"},{"url":"https://github.com/morcoan/JevSeek/blob/030d59114a03720783e839481a7814e7aecaad85/README.md","note":"README.md"}],"pinned":false,"topics":["coding-agent","deepseek","llm-agents","mcp","python","pywebview","react"]},{"id":"hev:reranker","name":"reranker","author":"hev","url":"https://github.com/hev/reranker","category":"Data & Search","plainSummary":"把查询和最多约 30 篇候选放进一次 Jev state，每篇一个 Noul「是否相关」，用来过滤或重排。","plainSummaryEn":"A Python reranker that packs a query and up to 30 candidates into one Jev state, with one Noul relevance question per document.","jevDecisionPoint":"每篇文档一个 Noul 相关度；超长列表再切批并发请求，结果可按阈值丢掉或按分数排序。","jevDecisionPointEn":"One Noul relevance question per document; long lists are chunked and scored, then filtered by threshold or sorted.","highlightBenefit":"用概率当阈值或排序键，不必再接一个生成式重排器。","highlightBenefitEn":"Uses a probability as a threshold or sort key instead of a generative reranker.","tags":["search-retrieval","classification-ranking"],"stars":9,"forks":0,"openIssues":1,"license":"Apache-2.0","language":"Python","lastCommitAt":"2026-09-17T20:47:19Z","createdAt":"2026-09-17T12:39:43Z","metadataFetchedAt":"2026-09-21T22:07:03.032Z","avatarUrl":"https://avatars.githubusercontent.com/u/20188639?v=4","summarySource":"curated","claimStatus":"源码已核验。支持 BM25 粗排后重排流水线与线程池并发；实际 BEIR 评测集收益需根据具体文档领域评测。","claimStatusEn":"Source code verified against pinned commit. Includes pipeline examples and concurrency handling; retrieval gains vary across datasets.","plainSummaryJa":"クエリと最大約 30 件の候補を 1 つの Jev state に入れ、文書ごとに Noul「関連するか」を問い、フィルタまたはリランクに使う。","plainSummaryKo":"질의와 최대 약 30개 후보를 하나의 Jev state에 넣고, 문서마다 Noul 관련 여부를 물어 필터 또는 재순위에 쓴다.","jevDecisionPointJa":"文書ごとに 1 つの Noul 関連度。長いリストは分割して並行リクエストし、しきい値で落とすか点数で並べる。","jevDecisionPointKo":"문서마다 Noul 관련도 하나. 긴 목록은 나눠 병렬 요청한 뒤 임계값으로 버리거나 점수로 정렬한다.","highlightBenefitJa":"確率をしきい値または並び替えキーとして使い、生成型リランカーを足さなくてよい。","highlightBenefitKo":"확률을 임계값이나 정렬 키로 써서 생성형 리랭커를 따로 붙이지 않아도 된다.","claimStatusJa":"固定コミットのソースを確認済み。BM25 連携や並行処理を含むが、実際の検索精度向上幅はデータセットに依存します。","claimStatusKo":"고정 커밋 기준 소스 확인 완료. BM25 연동 및 스레드 병렬 처리를 포함하나 검색 품질 개선 폭은 데이터셋에 따라 다릅니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T13:00:00.000000+00:00","evidence":[{"url":"https://github.com/hev/reranker/blob/1eb47266270b32b2a3667f9fb89646378ca9c9d6/hev_rerank/rerank.py#L1-L105","note":"hev_rerank/rerank.py (L1-L105)"},{"url":"https://github.com/hev/reranker/blob/1eb47266270b32b2a3667f9fb89646378ca9c9d6/examples/bm25_then_jev.py","note":"examples/bm25_then_jev.py (L1-L16)"},{"url":"https://github.com/hev/reranker/blob/1eb47266270b32b2a3667f9fb89646378ca9c9d6/hev_rerank/__init__.py","note":"hev_rerank/__init__.py (L1-L5)"},{"url":"https://github.com/hev/reranker/blob/1eb47266270b32b2a3667f9fb89646378ca9c9d6/README.md","note":"README.md"}],"pinned":false},{"id":"bunsdev:typesafe-ai-playground","name":"typesafe-ai-playground","author":"BunsDev","url":"https://github.com/TypeSafeAI/typesafe-playground","category":"Domain & Vertical Tools","plainSummary":"TypeSafe AI Jev 社区试验场，内置 110 个分类、对话路由、提取与决策实验用例，支持移动端交互与 A/B 对比。","plainSummaryEn":"Community playground for TypeSafe AI and Jev featuring 110 real-world scenarios, dilemmas, and interactive experiments.","jevDecisionPoint":"在 Next.js 服务端路由中向 Jev 发起状态与提问负载，实时展示离散判定分布、推理耗时与置信度。","jevDecisionPointEn":"Submits custom states and questions to Jev via Next.js API routes, displaying discrete probabilities, latency, and confidence scores.","highlightBenefit":"提供开箱即用的 Web 可视化界面，方便开发者直观调试 Jev 提示词、观察决策原语边界并对比不同模型表现。","highlightBenefitEn":"Offers an intuitive web interface for exploring prompt criteria, testing edge cases, and evaluating discrete decision outputs.","tags":["evaluation-benchmarks","typed-decisions"],"stars":18,"forks":5,"openIssues":1,"license":"MIT","language":"TypeScript","lastCommitAt":"2026-09-21T06:54:47Z","createdAt":"2026-09-16T16:20:23Z","metadataFetchedAt":"2026-09-21T22:07:06.044Z","avatarUrl":"https://avatars.githubusercontent.com/u/331025984?v=4","summarySource":"human-reviewed","claimStatus":"源码已核验。包含完整的 UI 组件与上百个内置用例；在线演示的响应速度取决于部署端与远程接口连通性。","claimStatusEn":"Source code verified against pinned commit. Complete Next.js application codebase; live demo latency depends on deployment hosting.","plainSummaryJa":"TypeSafe AI および Jev 向けのコミュニティプレイグラウンドであり、110 種類の実践的な分類や実験シナリオを Web 上で試すことができます。","plainSummaryKo":"TypeSafe AI 및 Jev를 위한 커뮤니티 플레이그라운드로 110개 실전 시나리오와 인터랙티브 실험을 제공합니다.","jevDecisionPointJa":"Next.js API ルート経由で Jev に状態と質問を送信し、離散確率分布や応答速度、確信度をリアルタイム表示します。","jevDecisionPointKo":"Next.js API 라우트를 통해 Jev에 상태와 질문을 전송하고 이산 확률 분포와 응답 지연 시간, 신뢰도를 표시합니다.","highlightBenefitJa":"直感的な Web インターフェースを通じて、プロンプト条件のチューニングや判断精度の事前検証が容易になります。","highlightBenefitKo":"직관적인 웹 인터페이스를 통해 프롬프트 기준을 조율하고 경계 조건을 시각적으로 확인하며 실험할 수 있습니다.","claimStatusJa":"固定コミットのソースを確認済み。豊富な UI とサンプル集を含むが、実際のレスポンス時間はホスティング環境に依存します。","claimStatusKo":"고정 커밋 기준 소스 확인 완료. 전체 애플리케이션 코드를 확인했으나 실시간 지연 시간은 배포 환경에 따라 달라집니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T13:00:00.000000+00:00","evidence":[{"url":"https://github.com/BunsDev/typesafe-ai-playground/blob/f67c3571fca1c468b4585f17282a58a1f57b1a17/app/api/run/route.ts#L1-L65","note":"app/api/run/route.ts (L1-L65)"},{"url":"https://github.com/TypeSafeAI/typesafe-playground/blob/f67c3571fca1c468b4585f17282a58a1f57b1a17/README.md","note":"README.md"}],"pinned":false},{"id":"zhengxuyu:litjev","name":"litjev","author":"zhengxuyu","url":"https://github.com/zhengxuyu/litjev","category":"High-Frequency & Simulation","plainSummary":"将开源大模型转化为 Jev 决策层的开放复现实现，基于 Qwen 等模型直接读取选项 logits 提供 System One 兼容接口。","plainSummaryEn":"Open reproduction of Jev decision layer on open-weights models, serving the System One schema by inspecting option logits directly.","jevDecisionPoint":"复刻 Jev 的 /v1/systemone 协议（Choice、Score、Noul），通过本地模型 forward 计算选项 token 的相对概率而非生成文本。","jevDecisionPointEn":"Implements the /v1/systemone endpoint (Choice, Score, Noul) by calculating candidate token log-probabilities without text generation.","highlightBenefit":"让开发者可在本地私有部署 Jev 兼容的离散决策服务，摆脱闭源商用 API 依赖并支持 MMLU-Pro 等评测。","highlightBenefitEn":"Enables private and local hosting of Jev-compatible decision services on GPUs without relying on proprietary cloud APIs.","tags":["evaluation-benchmarks","typed-decisions"],"stars":36,"forks":6,"openIssues":3,"license":"Apache-2.0","language":"Python","lastCommitAt":"2026-09-21T15:10:50Z","createdAt":"2026-09-17T17:19:25Z","metadataFetchedAt":"2026-09-21T22:07:09.021Z","avatarUrl":"https://avatars.githubusercontent.com/u/20417640?v=4","summarySource":"human-reviewed","claimStatus":"源码已核验。实现了一致的 API 接口与 logits 计算；在特定任务上的校准概率与原始 Jev 模型存在分布差异。","claimStatusEn":"Source code verified against pinned commit. Complete local API server; calibrated probability distribution may differ from hosted Jev.","plainSummaryJa":"オープンソースモデルを Jev 互換の決定層に変換するオープン再現実装であり、選択肢の logits から System One API を提供します。","plainSummaryKo":"오픈소스 모델을 Jev 호환 판단 레이어로 변환하는 오픈 구현체로 토큰 로짓을 읽어 System One 규격을 제공합니다.","jevDecisionPointJa":"Choice、Score、Noul のスキーマを実装し、テキスト生成を行わずに候補トークンの log 確率から直接判定を計算します。","jevDecisionPointKo":"Choice, Score, Noul 규격을 구현하여 텍스트 생성 없이 후보 토큰의 확률을 계산하여 즉시 반환합니다.","highlightBenefitJa":"独自の GPU 環境で Jev 互換の離散判定サービスをセルフホストでき、閉域環境での運用を可能にします。","highlightBenefitKo":"독자적인 GPU 환경에서 Jev 호환 이산 판단 서비스를 구축하여 폐쇄망 및 로컬 환경에서 활용할 수 있습니다.","claimStatusJa":"固定コミットのソースを確認済み。ローカル推論サーバーを含むが、較正された確率分布は本家 Jev と異なる場合があります。","claimStatusKo":"고정 커밋 기준 소스 확인 완료. 로컬 추론 API를 포함하나 보정된 확률 분포는 공식 Jev와 차이가 있을 수 있습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T13:00:00.000000+00:00","evidence":[{"url":"https://github.com/zhengxuyu/litjev/blob/f21216c9fe5afe7fa52ff7064a402ee57fdbddd3/src/litjev/api.py#L1-L83","note":"src/litjev/api.py (L1-L83)"},{"url":"https://github.com/zhengxuyu/litjev/blob/f21216c9fe5afe7fa52ff7064a402ee57fdbddd3/README.md","note":"README.md"}],"pinned":false},{"id":"mrnugget:jev-shell-history","name":"jev-shell-history","author":"mrnugget","url":"https://github.com/mrnugget/jev-shell-history","category":"CLI & Pipelines","plainSummary":"类似于 Fish 终端样式的 Zsh 历史命令建议工具，利用 Jev 对已有历史记录根据当前上下文进行智能打分排序。","plainSummaryEn":"Fish-style Zsh history suggestion tool ranked by Jev, ordering candidate commands from local history based on context.","jevDecisionPoint":"将当前敲入的命令前缀与本地 Zsh 历史候选组装为 Jev 请求，由 Jev 评估最佳补全项，仅做行内建议而不自动执行。","jevDecisionPointEn":"Packages current command input and historical candidates into a Jev request, ranking the most plausible command for inline suggestion.","highlightBenefit":"结合了确定性本地历史的安全性与 Jev 上下文语义理解能力，避免盲目基于字符串前缀匹配造成的低质建议。","highlightBenefitEn":"Combines the safety of local shell history with Jev semantic relevance, filtering out stale or risky command candidates.","tags":["cli-git-gates","classification-ranking"],"stars":94,"forks":6,"openIssues":3,"license":null,"language":"TypeScript","lastCommitAt":"2026-09-18T08:10:57Z","createdAt":"2026-09-18T08:11:07Z","metadataFetchedAt":"2026-09-21T22:07:12.026Z","avatarUrl":"https://avatars.githubusercontent.com/u/1185253?v=4","summarySource":"human-reviewed","claimStatus":"源码已核验。支持建议打分与终端管道交互；网络延迟可能影响单字符按键响应，建议搭配本地缓存使用。","claimStatusEn":"Source code verified against pinned commit. Terminal hook and suggestion logic inspected; interactive typing requires fast network connectivity.","plainSummaryJa":"Jev を利用して Zsh の履歴候補を文脈に合わせてランク付けし、Fish スタイルのインライン補完を表示する CLI ツールです。","plainSummaryKo":"Jev를 사용하여 Zsh 로컬 기록 명령어를 문맥에 맞게 순위 매겨 인라인 제안을 제공하는 터미널 도구입니다.","jevDecisionPointJa":"現在の入力とローカル履歴を Jev に送信し、最も適切なコマンド候補を評価してインライン表示します（自動実行はしません）。","jevDecisionPointKo":"현재 입력 중인 명령어와 로컬 히스토리를 Jev로 평가하여 가장 타당한 완성 후보를 선정합니다 (자동 실행 안 됨).","highlightBenefitJa":"文字列前方一致にとどまらず、文脈に即した適切なコマンドを優先提示してシェル操作を高速化します。","highlightBenefitKo":"단순 접두사 매칭을 넘어 문맥에 부합하는 명령어를 우선 노출하여 터미널 작업 생산성을 높입니다.","claimStatusJa":"固定コミットのソースを確認済み。シェル連携スクリプトを確認。快適な入力には低遅延なネットワーク環境が推奨されます。","claimStatusKo":"고정 커밋 기준 소스 확인 완료. 셸 연동 및 제안 로직을 확인했으나 타자 반응성은 네트워크 지연에 영향을 받습니다.","licenseStatus":"unconfirmed","sourceReviewedAt":"2026-09-19T13:00:00.000000+00:00","evidence":[{"url":"https://github.com/mrnugget/jev-shell-history/blob/4b2b75d26c0ccf5726263904514a22a8e11659ea/src/suggest.ts#L1-L196","note":"src/suggest.ts (L1-L196)"},{"url":"https://github.com/mrnugget/jev-shell-history/blob/4b2b75d26c0ccf5726263904514a22a8e11659ea/README.md","note":"README.md"}],"pinned":false},{"id":"luantak:is-malicious","name":"is-malicious","author":"luantak","url":"https://github.com/luantak/is-malicious","category":"Security & Guardrails","plainSummary":"命令行代码库恶意行为扫描器，在运行未知代码前利用 Jev 分析源码、CI 配置与构建脚本的可疑行为。","plainSummaryEn":"CLI security scanner using Jev to inspect source files, build scripts, and CI configs for suspicious patterns before execution.","jevDecisionPoint":"将待检文件内容切片与安全规则输入 Jev，由 Jev 判断代码是否包含数据外发、混淆加载或凭据嗅探等风险。","jevDecisionPointEn":"Feeds file contents and security criteria into Jev to assess risks such as data exfiltration, obfuscated loaders, and credential leakage.","highlightBenefit":"在不直接执行脚本的安全沙箱前置阶段提供多维度危险信号预警，帮助开发者排查供应链中毒隐患。","highlightBenefitEn":"Provides rapid pre-execution safety screening across multiple file formats without needing a heavy static analysis toolchain.","tags":["security-guardrails","cli-git-gates"],"stars":21,"forks":3,"openIssues":1,"license":"MIT","language":"TypeScript","lastCommitAt":"2026-09-21T08:55:20Z","createdAt":"2026-09-18T11:40:17Z","metadataFetchedAt":"2026-09-21T22:07:15.033Z","avatarUrl":"https://avatars.githubusercontent.com/u/46541492?v=4","summarySource":"human-reviewed","claimStatus":"源码已核验。包含检测规则与行号关联输出；无法替代确定性 AST 静态分析与动态沙箱，需作为辅助防御层。","claimStatusEn":"Source code verified against pinned commit. Pattern rules and CLI outputs verified; serves as an advisory layer alongside deterministic linters.","plainSummaryJa":"未知のコードを実行する前に、Jev を使用してソースファイルや CI スクリプトの不審な挙動を検知する CLI セキュリティスキャナーです。","plainSummaryKo":"실행 전 소스 코드, 빌드 스크립트, CI 설정에서 의심스러운 동작을 Jev로 탐지하는 CLI 보안 스캐너입니다.","jevDecisionPointJa":"ファイル内容とセキュリティ判定基準を Jev に渡し、情報漏洩や難読化コード、認証情報の不正取得リスクを評価します。","jevDecisionPointKo":"파일 내용과 보안 기준을 Jev에 입력하여 데이터 유출, 난독화 로더, 자격 증명 탈취 등의 위험성을 판별합니다.","highlightBenefitJa":"重い静的解析ツールをセットアップすることなく、実行前の迅速なセキュリティリスク評価を提供します。","highlightBenefitKo":"무거운 정적 분석 도구 없이도 실행 전 단계에서 신속하게 다각도 위험 징후를 감지할 수 있습니다.","claimStatusJa":"固定コミットのソースを確認済み。ルール定義と CLI 出力を確認。確定的な静的解析と併用する補助層として機能します。","claimStatusKo":"고정 커밋 기준 소스 확인 완료. 탐지 규칙 및 CLI 출력을 확인했으나 결정론적 정적 분석과 병행해야 합니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T13:00:00.000000+00:00","evidence":[{"url":"https://github.com/luantak/is-malicious/blob/b6052465eb47cf85f347596385954aa06532718a/src/jev.ts#L1-L181","note":"src/jev.ts (L1-L181)"},{"url":"https://github.com/luantak/is-malicious/blob/b6052465eb47cf85f347596385954aa06532718a/README.md","note":"README.md"}],"pinned":false,"topics":["jev","security-scanner","typesafe-ai"]},{"id":"gargpratyush:jev-router","name":"jev-router","author":"gargpratyush","url":"https://github.com/gargpratyush/jev-router","category":"Routing & Cost Optimization","plainSummary":"Claude Code / CLI 代理：Jev 给任务复杂度打分，并在当前账号可用的模型里选一个，再由本地策略决定是否更换。","plainSummaryEn":"A Claude Code and CLI proxy that asks Jev to score task complexity and pick a model from the account’s available set.","jevDecisionPoint":"三个 Score 衡量任务、推理和工具复杂度，再在可用模型上做 Choice；置信不足或 Jev 失败则保持当前模型。","jevDecisionPointEn":"Scores task, reasoning, and tool complexity, then chooses among the exact models the account can run; failures keep the current model.","highlightBenefit":"把模型选择做成可检查的闭集判断；Jev 不可用时不打断当前会话。","highlightBenefitEn":"Makes model choice an inspectable closed-set decision, and never blocks a prompt when Jev is down.","tags":["llm-routing-cost","coding-agents"],"stars":309,"forks":30,"openIssues":22,"license":"MIT","language":"JavaScript","lastCommitAt":"2026-09-19T08:37:38Z","createdAt":"2026-09-16T20:39:22Z","metadataFetchedAt":"2026-09-21T22:07:18.010Z","avatarUrl":"https://avatars.githubusercontent.com/u/75370041?v=4","summarySource":"curated","claimStatus":"源码已核验。包含 Claude Code 包装脚本、策略配置与状态栏展示；实际节约比例取决于用户日常提示词复杂度分布。","claimStatusEn":"Source code verified against pinned commit. Includes wrapper binaries and policy configs; cost savings vary with workload distribution.","plainSummaryJa":"Claude Code / CLI のプロキシ。Jev がタスクの複雑さを採点し、アカウントで使えるモデルから 1 つ選び、ローカル方針が切り替えを決める。","plainSummaryKo":"Claude Code / CLI 프록시. Jev가 작업 복잡도를 매기고 계정에서 쓸 수 있는 모델 중 하나를 고른 뒤, 로컬 정책이 교체 여부를 정한다.","jevDecisionPointJa":"3 つの Score でタスク・推論・ツールの複雑さを測り、使えるモデル上で Choice する。信頼が足りない、または Jev が失敗したら今のモデルを維持する。","jevDecisionPointKo":"Score 셋으로 작업·추론·도구 복잡도를 재고 사용 가능한 모델에서 Choice한다. 신뢰가 낮거나 Jev가 실패하면 현재 모델을 유지한다.","highlightBenefitJa":"モデル選択を検査できる閉集合の判断にする。Jev が使えないときも今のセッションを止めない。","highlightBenefitKo":"모델 선택을 검사 가능한 닫힌 집합 판단으로 만든다. Jev를 쓸 수 없어도 현재 세션을 막지 않는다.","claimStatusJa":"固定コミットのソースを確認済み。ラッパースクリプトとルーティング設定を確認。削減効果は日常のタスク傾向に依存します。","claimStatusKo":"고정 커밋 기준 소스 확인 완료. 래퍼 스크립트와 정책 설정을 확인했으나 실제 비용 절감률은 작업 분포에 따라 다릅니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T13:00:00.000000+00:00","evidence":[{"url":"https://github.com/gargpratyush/jev-router/blob/38da6b84ea01241bfc41fbddc0928d0f40a703f0/src/policy.mjs#L1-L63","note":"src/policy.mjs (L1-L63)"},{"url":"https://github.com/gargpratyush/jev-router/blob/38da6b84ea01241bfc41fbddc0928d0f40a703f0/bin/jev-claude.mjs","note":"bin/jev-claude.mjs (L1-L157)"},{"url":"https://github.com/gargpratyush/jev-router/blob/38da6b84ea01241bfc41fbddc0928d0f40a703f0/bin/jev-codex.mjs","note":"bin/jev-codex.mjs (L1-L5)"},{"url":"https://github.com/gargpratyush/jev-router/blob/38da6b84ea01241bfc41fbddc0928d0f40a703f0/bin/jev-explain.mjs","note":"bin/jev-explain.mjs (L1-L6)"},{"url":"https://github.com/gargpratyush/jev-router/blob/38da6b84ea01241bfc41fbddc0928d0f40a703f0/bin/jev-statusline.mjs","note":"bin/jev-statusline.mjs (L1-L44)"},{"url":"https://github.com/gargpratyush/jev-router/blob/38da6b84ea01241bfc41fbddc0928d0f40a703f0/README.md","note":"README.md"}],"pinned":false},{"id":"aaronshaf:opencode-jev-orchestrator","name":"opencode-jev-orchestrator","author":"aaronshaf","url":"https://github.com/aaronshaf/opencode-jev-orchestrator","category":"Routing & Cost Optimization","plainSummary":"OpenCode 编排器：会话停在廉价父模型上，Jev 判定本轮偏难时才通过工具拉起更强的子 Agent。","plainSummaryEn":"An OpenCode orchestrator that keeps a cheap sticky parent model and, when Jev flags a hard turn, escalates through a child subagent.","jevDecisionPoint":"三个 Score 衡量任务、推理和工具复杂度，再 Choice 选出 fast / balanced / strong / long；本地策略决定停留、升级或并行。","jevDecisionPointEn":"Scores task, reasoning, and tool complexity, then chooses fast / balanced / strong / long; local policy stays, escalates, or fans out.","highlightBenefit":"会话停在廉价父模型上；只有判定为难的轮次才另开更强的子 Agent。","highlightBenefitEn":"Keeps a cheap sticky parent model and opens a stronger child only when Jev flags a hard turn.","tags":["llm-routing-cost","coding-agents"],"stars":2,"forks":0,"openIssues":0,"license":"MIT","language":"TypeScript","lastCommitAt":"2026-09-18T21:12:53Z","createdAt":"2026-09-18T13:56:30Z","metadataFetchedAt":"2026-09-21T22:07:21.051Z","avatarUrl":"https://avatars.githubusercontent.com/u/122654?v=4","summarySource":"curated","claimStatus":"源码已核验。包含决策提取与子智能体派发逻辑；子任务上下文隔离与合并质量取决于任务切分边界。","claimStatusEn":"Source code verified against pinned commit. Core routing and subagent dispatching verified; synthesis quality depends on task boundaries.","plainSummaryJa":"OpenCode のオーケストレータ。セッションは安い親モデルに留め、Jev がこのターンを難しいと判定したときだけツール経由で強い子 Agent を立てる。","plainSummaryKo":"OpenCode 오케스트레이터. 세션은 저비용 부모 모델에 두고, Jev가 이번 턴을 어렵다고 볼 때만 도구로 더 강한 자식 Agent를 띄운다.","jevDecisionPointJa":"3 つの Score でタスク・推論・ツールの複雑さを測り、Choice で fast / balanced / strong / long を選ぶ。ローカル方針が残留・エスカレーション・並列を決める。","jevDecisionPointKo":"Score 셋으로 작업·추론·도구 복잡도를 재고 Choice로 fast / balanced / strong / long을 고른다. 로컬 정책이 유지, 승격, 병렬을 정한다.","highlightBenefitJa":"セッションは安い親モデルに留める。難しいと判定されたターンだけ、より強い子 Agent を別途立てる。","highlightBenefitKo":"세션은 저비용 부모 모델에 둔다. 어렵다고 판정된 턴만 더 강한 자식 Agent를 따로 연다.","claimStatusJa":"固定コミットのソースを確認済み。ルーティングおよびサブエージェント制御を確認。統合精度はタスク分割の粒度に依存します。","claimStatusKo":"고정 커밋 기준 소스 확인 완료. 라우팅 및 서브 에이전트 위임 로직을 확인했으나 분할 경계 설정에 영향을 받습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T13:00:00.000000+00:00","evidence":[{"url":"https://github.com/aaronshaf/opencode-jev-orchestrator/blob/ffe76fed66b3d544f218cacff1947fc191688c3a/src/jev.ts#L1-L193","note":"src/jev.ts (L1-L193)"},{"url":"https://github.com/aaronshaf/opencode-jev-orchestrator/blob/ffe76fed66b3d544f218cacff1947fc191688c3a/README.md","note":"README.md"}],"pinned":false},{"id":"leonaaardob:fast-dev-compaction","name":"fast-dev-compaction","author":"leonaaardob","url":"https://github.com/leonaaardob/fast-dev-compaction","category":"Context GC & Filter","plainSummary":"Codex 插件与上下文压缩工具，在会话生命周期钩子中利用 Jev 判定历史记录的保留价值并进行无损还原。","plainSummaryEn":"Codex plugin for context compaction, using Jev within session lifecycle hooks to preserve essential history while pruning noise.","jevDecisionPoint":"在会话上下文触达上限时，由 Jev 逐条评估历史消息和工具调用的保留必要性，仅剪除冗余噪音。","jevDecisionPointEn":"Assesses past conversation turns with Jev when context limits approach, selecting critical tool outputs and user requirements for retention.","highlightBenefit":"相比简单的截断或整段摘要，通过离散判断保留了关键的代码定位与协议结构，防止长程编码对话失忆。","highlightBenefitEn":"Avoids blanket truncation or lossy re-summarization by selectively preserving exact code locations and decision traces.","tags":["context-compaction","coding-agents"],"stars":3,"forks":0,"openIssues":1,"license":"MIT","language":"TypeScript","lastCommitAt":"2026-09-19T01:03:52Z","createdAt":"2026-09-18T00:35:35Z","metadataFetchedAt":"2026-09-21T22:07:24.019Z","avatarUrl":"https://avatars.githubusercontent.com/u/207195840?v=4","summarySource":"human-reviewed","claimStatus":"源码已核验。基于 tamaratran/fast-jev-compaction 移植到 Codex 钩子；实际压缩比与还原完整性因提示词内容而异。","claimStatusEn":"Source code verified against pinned commit. Adapted for Codex lifecycle hooks; compaction ratios depend on conversation style.","plainSummaryJa":"セッションフック内で Jev を活用し、必要な履歴情報を保持しながら不要なログを間引く Codex 向けコンテキスト圧縮プラグインです。","plainSummaryKo":"세션 수명주기 훅에서 Jev를 호출하여 핵심 기록을 보존하고 불필요한 로그를 정리하는 Codex용 컨텍스트 압축 플러그인입니다.","jevDecisionPointJa":"コンテキスト長の上限に近づいた際、Jev が各履歴やツール出力の重要度を判定し、重要な情報のみを残して圧縮します。","jevDecisionPointKo":"컨텍스트 한도 도달 시 Jev가 각 턴과 도구 호출 결과의 보존 가치를 판단하여 중요한 단서만 선별적으로 유지합니다.","highlightBenefitJa":"単純な切り捨てや過度な要約による情報の欠落を防ぎ、長時間のコーディングセッションでも文脈を維持できます。","highlightBenefitKo":"단순 절삭이나 손실성 요약 대신 정확한 코드 위치와 결정 경로를 보존하여 긴 코딩 세션에서도 문맥을 유지합니다.","claimStatusJa":"固定コミットのソースを確認済み。Codex フックへの移植実装を確認。実際の圧縮効率は会話履歴の内容に依存します。","claimStatusKo":"고정 커밋 기준 소스 확인 완료. Codex 라이프사이클 훅 연동을 확인했으나 실제 압축률은 대화 형식에 따라 달라집니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T13:00:00.000000+00:00","evidence":[{"url":"https://github.com/leonaaardob/fast-dev-compaction/blob/7147a4bab4ff9caf64ae44a2db1ab1f907a20291/dist/compact.js#L1-L234","note":"dist/compact.js (L1-L234)"},{"url":"https://github.com/leonaaardob/fast-dev-compaction/blob/7147a4bab4ff9caf64ae44a2db1ab1f907a20291/README.md","note":"README.md"}],"pinned":false,"topics":["codex","codex-plugin","compaction","jev","typesafe"]},{"id":"compozy:yoshi","name":"yoshi","author":"compozy","url":"https://github.com/compozy/yoshi","category":"Context GC & Filter","plainSummary":"面向 Claude Code 与 Codex 的上下文剪枝代理，通过 Jev 评估历史条目必要性并保持工具调用协议结构完整。","plainSummaryEn":"Context-pruning proxy for Claude Code and Codex, employing Jev to prune redundant history while keeping tool protocols intact.","jevDecisionPoint":"在代理转发层使用 Jev 对消息历史打分，过滤已失效的中间试错输出，仅向大模型提交精简后的有效上下文。","jevDecisionPointEn":"Intercepts request payloads and uses Jev to score historical messages, discarding outdated tool noise before forwarding to upstream LLMs.","highlightBenefit":"降低每次交互的输入 Token 数量，加快响应首字时间，且不破坏现有客户端与服务端的协议兼容性。","highlightBenefitEn":"Decreases input token counts and time-to-first-token while preserving protocol compliance with client tool calling standards.","tags":["context-compaction","coding-agents"],"stars":21,"forks":2,"openIssues":0,"license":"MIT","language":"TypeScript","lastCommitAt":"2026-09-18T02:46:28Z","createdAt":"2026-09-18T00:02:20Z","metadataFetchedAt":"2026-09-21T22:07:27.059Z","avatarUrl":"https://avatars.githubusercontent.com/u/205107042?v=4","summarySource":"human-reviewed","claimStatus":"源码已核验。当前为概念验证阶段（POC），包含校准基准脚本；更广泛的生产集成需关注上游项目进展。","claimStatusEn":"Source code verified against pinned commit. Proof-of-concept with calibration benchmarks; upstream library integration ongoing.","plainSummaryJa":"Claude Code や Codex 向けのコンテキスト削減プロキシであり、Jev を使って履歴の必要性を評価しツールプロトコルを維持したまま剪定します。","plainSummaryKo":"도구 호출 프로토콜을 보존하면서 Jev로 불필요한 기록을 정리하는 Claude Code 및 Codex용 컨텍스트 정리 프록시입니다.","jevDecisionPointJa":"プロキシ層でリクエストをインターセプトし、Jev が不要と判断した試行錯誤ログを除去してから上位モデルへ転送します。","jevDecisionPointKo":"프록시 계층에서 요청을 가로채 Jev로 메시지 중요도를 평가한 뒤 유효하지 않은 중간 출력을 제거하고 전달합니다.","highlightBenefitJa":"入力トークン数と初期応答時間を削減しつつ、既存クライアントのツール呼び出しプロトコルとの完全な互換性を保ちます。","highlightBenefitKo":"입력 토큰 수와 첫 토큰 응답 시간을 줄이면서도 도구 호출 프로토콜의 표준 규격을 유지합니다.","claimStatusJa":"固定コミットのソースを確認済み。較正ベンチマークを含む検証用実装であり、本番運用にはさらなる統合が必要です。","claimStatusKo":"고정 커밋 기준 소스 확인 완료. 보정 벤치마크를 포함한 개념 검증 구현이며 상용 배포 시 추가 통합이 필요합니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T13:00:00.000000+00:00","evidence":[{"url":"https://github.com/compozy/yoshi/blob/55c719718e5039f2276fbad804211682ae012f1e/benchmarks/jev-calibrate.ts#L1-L188","note":"benchmarks/jev-calibrate.ts (L1-L188)"},{"url":"https://github.com/compozy/yoshi/blob/55c719718e5039f2276fbad804211682ae012f1e/README.md","note":"README.md"}],"pinned":false,"topics":["bun","claude-code","codex","context-window","jev","llm-proxy","typesafe"]},{"id":"iamvatsalpatel:tiershift","name":"tiershift","author":"iamvatsalpatel","url":"https://github.com/iamvatsalpatel/tiershift","category":"Routing & Cost Optimization","plainSummary":"基于 YAML 声明策略的模型分级路由工具，在约 180 毫秒内通过 Jev 将请求路由到满足要求的最低成本模型。","plainSummaryEn":"Policy-driven model routing framework routing every LLM call to the cheapest capable tier in ~180 ms via TypeSafe Jev.","jevDecisionPoint":"无需微调训练数据，利用纯文本 YAML 规则由 Jev 对请求进行离散分类判断，命中阈值即向下分流。","jevDecisionPointEn":"Uses declarative YAML policies evaluated by Jev discrete classification without requiring pre-collected training data.","highlightBenefit":"支持 TypeScript 与 Python 双语言生态，极低决策延迟，帮助团队透明化控制大模型调用成本。","highlightBenefitEn":"Provides unified TypeScript and Python interfaces to minimize inference spend with minimal routing overhead.","tags":["llm-routing-cost","typed-decisions"],"stars":3,"forks":1,"openIssues":0,"license":"MIT","language":"TypeScript","lastCommitAt":"2026-09-18T05:36:33Z","createdAt":"2026-09-17T20:26:56Z","metadataFetchedAt":"2026-09-21T22:07:30.053Z","avatarUrl":"https://avatars.githubusercontent.com/u/61371035?v=4","summarySource":"human-reviewed","claimStatus":"源码已核验。包含门禁实验脚本与双语言实现；实际端到端耗时与网络延迟及各服务商接口波动相关。","claimStatusEn":"Source code verified against pinned commit. Includes gate benchmarks and dual-language tests; latency depends on regional network proximity.","plainSummaryJa":"YAML ポリシーに基づいて Jev が約 180ms でリクエストを判定し、要件を満たす最も安価なモデル層へ振り分けるルーティング基盤です。","plainSummaryKo":"YAML 정책을 바탕으로 Jev가 약 180ms 내에 요청을 판별하여 가장 저렴한 적합 모델 계층으로 전달하는 라우팅 프레임워크입니다.","jevDecisionPointJa":"事前学習データ不要で、プレーンな YAML 基準を Jev に評価させてリクエストの複雑度に応じた振り分けを実行します。","jevDecisionPointKo":"별도의 훈련 데이터 없이 YAML 규칙에 따라 Jev가 이산 분류를 수행하여 요청 복잡도에 맞게 분기합니다.","highlightBenefitJa":"TypeScript と Python の双方に対応し、最小限の判定遅延で推論コストの最適化を実現します。","highlightBenefitKo":"TypeScript와 Python 생태계를 모두 지원하며 짧은 라우팅 지연 시간으로 모델 사용 비용을 체계적으로 절감합니다.","claimStatusJa":"固定コミットのソースを確認済み。実験ベンチマークと両言語のコードを確認。実環境での遅延は接続状況に依存します。","claimStatusKo":"고정 커밋 기준 소스 확인 완료. 벤치마크 및 양 언어 테스트를 확인했으나 실제 통신 지연 시간은 네트워크 환경에 따라 달라집니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T13:00:00.000000+00:00","evidence":[{"url":"https://github.com/iamvatsalpatel/tiershift/blob/16a0826b9f62eba6699240c4b53a7237a8c41617/bench/experiments/gate-experiment.ts#L1-L83","note":"bench/experiments/gate-experiment.ts (L1-L83)"},{"url":"https://github.com/iamvatsalpatel/tiershift/blob/16a0826b9f62eba6699240c4b53a7237a8c41617/README.md","note":"README.md"}],"pinned":false,"topics":["agents","cost-optimization","jev","llm","model-routing","openai-compatible","proxy","python","typesafe","typescript"]},{"id":"abdelstark:heist-one","name":"heist-one","author":"AbdelStark","url":"https://github.com/AbdelStark/heist-one","category":"High-Frequency & Simulation","plainSummary":"可观察的浏览器潜行游戏，由 Jev 驱动守卫的强类型状态判断，而确定性代码引擎控制物理世界与移动规律。","plainSummaryEn":"Observable browser stealth game where Jev makes typed guard judgments while deterministic code owns the physics world.","jevDecisionPoint":"在游戏主循环或事件触发时，由 Jev 判定守卫的警觉状态（怀疑、警报、搜寻），并将决策输出实时同步给渲染引擎。","jevDecisionPointEn":"Queries Jev during gameplay ticks or events to determine guard alert states, feeding discrete decisions back into the game loop.","highlightBenefit":"探索了利用 Jev 作为游戏 NPC 离散决策大脑的可行性，使 AI 行为既具备语义理解力又受确定性游戏规则约束。","highlightBenefitEn":"Demonstrates using Jev for game AI behavior trees, achieving dynamic situational awareness within deterministic constraints.","tags":["games-simulation","typed-decisions"],"stars":7,"forks":0,"openIssues":2,"license":"MIT","language":"TypeScript","lastCommitAt":"2026-09-17T14:39:00Z","createdAt":"2026-09-17T07:29:14Z","metadataFetchedAt":"2026-09-21T22:07:33.088Z","avatarUrl":"https://avatars.githubusercontent.com/u/45264458?v=4","summarySource":"human-reviewed","claimStatus":"源码已核验。包含完整的游戏引擎与决策代码；频繁的远程接口调用在弱网环境下可能影响即时帧率。","claimStatusEn":"Source code verified against pinned commit. Complete monorepo game implementation; high-frequency remote calls require low network latency.","plainSummaryJa":"ブラウザー上で動作する潜入ゲームであり、確定的な物理エンジンと Jev によるガードの型付き状況判断を組み合わせています。","plainSummaryKo":"브라우저 잠입 게임으로 물리 법칙은 결정론적 코드가 제어하고 경비원의 상황 판단은 Jev의 타입화된 결정을 통해 수행됩니다.","jevDecisionPointJa":"ゲームループ内のトリガーに応じて Jev がガードの警戒レベルや行動方針を決定し、結果をレンダリング側に反映します。","jevDecisionPointKo":"게임 틱 또는 이벤트 발생 시 Jev를 호출하여 경비원의 경계 상태를 판정하고 결과를 게임 루프에 즉시 반영합니다.","highlightBenefitJa":"Jev をゲーム NPC の意思決定コアとして活用し、確定的なルールの中で自然かつ変化に富んだ敵の挙動を表現します。","highlightBenefitKo":"게임 NPC의 행동 트리에 Jev를 적용하여 규칙 기반 물리 시스템 안에서 유연하고 지능적인 상황 판단을 구현합니다.","claimStatusJa":"固定コミットのソースを確認済み。モノレポ構成のゲームコードを確認。高頻度呼び出し時のフレームレートは通信環境に依存します。","claimStatusKo":"고정 커밋 기준 소스 확인 완료. 모노레포 게임 전체 구현을 확인했으나 잦은 원격 호출 시 네트워크 상태가 프레임에 영향을 줄 수 있습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T13:00:00.000000+00:00","evidence":[{"url":"https://github.com/AbdelStark/heist-one/blob/632c9a55a1e5eb2cbf0b9f87db575f0b5eb36e8c/packages/game/src/decision.ts#L1-L451","note":"packages/game/src/decision.ts (L1-L451)"},{"url":"https://github.com/AbdelStark/heist-one/blob/632c9a55a1e5eb2cbf0b9f87db575f0b5eb36e8c/README.md","note":"README.md"}],"pinned":false,"topics":["ai-game","jev","phaser","probabilistic-ai","remotion","stealth-game","system-one-models","typesafe","typescript"]},{"id":"sorrycc:typesafe-snake","name":"typesafe-snake","author":"sorrycc","url":"https://github.com/sorrycc/typesafe-snake","category":"High-Frequency & Simulation","plainSummary":"由 TypeSafe Jev 模型自动操作的贪吃蛇游戏，每 tick 执行一次原子决策，合法移动与物理事实均由本地代码生成。","plainSummaryEn":"Snake autoplayer driven by TypeSafe Jev, executing one discrete System One decision per tick with code-enforced legal moves.","jevDecisionPoint":"将当前蛇身坐标、食物位置与合法转向候选组装为状态，由 Jev 选择下一步最佳转向（UP/DOWN/LEFT/RIGHT）。","jevDecisionPointEn":"Supplies board state, fruit location, and legal direction choices to Jev, selecting the optimal move direction each tick.","highlightBenefit":"验证了 Jev 在离散空间快速做二元/多元选择的单步决策能力，且底层严格杜绝撞墙等非法动作。","highlightBenefitEn":"Demonstrates fast discrete choice in constrained environments while deterministic logic prevents illegal collisions.","tags":["games-simulation","typed-decisions"],"stars":20,"forks":2,"openIssues":0,"license":null,"language":"TypeScript","lastCommitAt":"2026-09-17T09:42:33Z","createdAt":"2026-09-17T09:42:36Z","metadataFetchedAt":"2026-09-21T22:07:36.064Z","avatarUrl":"https://avatars.githubusercontent.com/u/35128?v=4","summarySource":"human-reviewed","claimStatus":"源码已核验。包含游戏引擎、Jev 客户端与提示词定义；单步网络往返延迟决定了贪吃蛇的最大平滑移动速度。","claimStatusEn":"Source code verified against pinned commit. Complete React/TypeScript implementation; tick frequency is bounded by network latency.","plainSummaryJa":"TypeSafe Jev が自動操作するスネークゲームであり、ティックごとに 1 回の離散判断を行い、移動可能判定はコード側で担保します。","plainSummaryKo":"TypeSafe Jev가 매 틱마다 이산 결정을 내려 스스로 플레이하는 스네이크 게임으로 유효한 이동 규칙은 코드가 보장합니다.","jevDecisionPointJa":"蛇の座標や餌の位置、進行可能な方向を状態として渡し、Jev が最適な次の方向（上下左右）を選択します。","jevDecisionPointKo":"뱀의 위치와 먹이 좌표, 이동 가능한 방향을 상태로 구성하여 Jev가 최적의 이동 방향을 선택하도록 합니다.","highlightBenefitJa":"制約された環境における Jev の高速選択能力を実証し、不正な移動は確定ロジックで排除してゲームの成立を保証します。","highlightBenefitKo":"제약된 환경에서 Jev의 단일 틱 이산 판단 능력을 검증하며 불법적인 충돌은 결정론적 코드가 방지합니다.","claimStatusJa":"固定コミットのソースを確認済み。React/TypeScript の実装を確認。ティック速度は API との通信往復遅延に依存します。","claimStatusKo":"고정 커밋 기준 소스 확인 완료. React 및 TypeScript 구현을 확인했으나 틱 진행 속도는 네트워크 왕복 시간에 좌우됩니다.","licenseStatus":"unconfirmed","sourceReviewedAt":"2026-09-19T13:00:00.000000+00:00","evidence":[{"url":"https://github.com/sorrycc/typesafe-snake/blob/8bf3f7c261ad35ece3345a02d21ba638ddfaf87f/src/jev/client.ts#L1-L29","note":"src/jev/client.ts (L1-L29)"},{"url":"https://github.com/sorrycc/typesafe-snake/blob/8bf3f7c261ad35ece3345a02d21ba638ddfaf87f/README.md","note":"README.md"}],"pinned":false},{"id":"harshil1712:slidepilot","name":"slidepilot","author":"harshil1712","url":"https://github.com/harshil1712/slidepilot","category":"Creative Tools","plainSummary":"基于语音语义理解的 Slidev 自动翻页控制器，运行于 Cloudflare Agents 与 Jev 之上。","plainSummaryEn":"Voice-driven semantic auto-advance controller for Slidev presentations powered by Cloudflare Agents and TypeSafe Jev.","jevDecisionPoint":"监听演讲者实时转录文本，由 Jev 评估当前页面核心要点是否已讲述完毕，并触发页面跳转或停顿。","jevDecisionPointEn":"Analyzes live speech transcripts with Jev to judge whether the current slide's topics are covered, triggering slide transitions.","highlightBenefit":"摆脱物理翻页笔束缚，结合确定性规则与语义完成度判定，实现自然流畅的演示文稿演讲同步。","highlightBenefitEn":"Enables hands-free slide navigation by combining semantic completion checks with manual safety overrides.","tags":["creative-multimedia","voice-conversation"],"stars":4,"forks":0,"openIssues":1,"license":"MIT","language":"TypeScript","lastCommitAt":"2026-09-18T09:44:08Z","createdAt":"2026-09-18T09:35:35Z","metadataFetchedAt":"2026-09-21T22:07:39.053Z","avatarUrl":"https://avatars.githubusercontent.com/u/18901032?v=4","summarySource":"human-reviewed","claimStatus":"源码已核验。包含 Cloudflare Worker 决策逻辑与测试；语音识别准确度与现场网络延迟会影响翻页时机。","claimStatusEn":"Source code verified against pinned commit. Includes Worker endpoints and test cases; transition timing depends on speech transcription fidelity.","plainSummaryJa":"Cloudflare Agents と Jev を組み合わせた Slidev 向け音声認識自動スライド送りコントローラーです。","plainSummaryKo":"Cloudflare Agents와 Jev를 기반으로 음성 전사 텍스트를 분석하여 자동으로 슬라이드를 넘겨주는 Slidev 프레젠테이션 컨트롤러입니다.","jevDecisionPointJa":"リアルタイムの音声書き起こしテキストを Jev で解析し、スライドの要点が十分に説明されたかを判定してページを進めます。","jevDecisionPointKo":"실시간 음성 전사 내용을 Jev로 평가하여 현재 슬라이드의 핵심 내용이 모두 다뤄졌는지 판단하고 전환을 트리거합니다.","highlightBenefitJa":"クリッカーを持たずにプレゼンを進行でき、意味論的な説明完了判定と手動オーバーライドを両立します。","highlightBenefitKo":"물리적인 리모컨 없이도 발표자의 음성 흐름에 맞춰 자연스럽게 슬라이드가 동기화되도록 지원합니다.","claimStatusJa":"固定コミットのソースを確認済み。Cloudflare Worker 上のロジックを確認。スライド送りの精度は音声認識の品質に影響されます。","claimStatusKo":"고정 커밋 기준 소스 확인 완료. Worker 엔드포인트 및 테스트를 확인했으나 음성 전사 품질과 지연 시간에 영향을 받습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T13:00:00.000000+00:00","evidence":[{"url":"https://github.com/harshil1712/slidepilot/blob/8ba4e89b56e4f9a08be2f1f69346240cab95a7b6/apps/worker/src/decision.ts#L1-L219","note":"apps/worker/src/decision.ts (L1-L219)"},{"url":"https://github.com/harshil1712/slidepilot/blob/8ba4e89b56e4f9a08be2f1f69346240cab95a7b6/README.md","note":"README.md"}],"pinned":false,"topics":["cloudflare-workers","jev","slidev","slidev-addon","typescript","voice","workers-ai"]},{"id":"trungdq88:youtube-sponsor-detection","name":"youtube-sponsor-detection","author":"trungdq88","url":"https://github.com/trungdq88/youtube-sponsor-detection","category":"Creative Tools","plainSummary":"结合实时音频与字幕由 Jev 驱动的 YouTube 视频赞助广告片段检测与自动跳过扩展。","plainSummaryEn":"YouTube extension detecting sponsored segments from live audio and transcripts using Jev, skipping promotional blocks automatically.","jevDecisionPoint":"提取视频字幕或音频转录切片，由 Jev 判定当前片段是否属于赞助商宣传读条，代码引擎控制时间戳与播放器快进。","jevDecisionPointEn":"Inspects transcript segments with Jev to identify sponsor reads while browser code handles playback timeline skipping.","highlightBenefit":"无需等待社区用户手动打点，借助 Jev 的语义判定能力实时识别个性化口播赞助，提升视频观看体验。","highlightBenefitEn":"Identifies promotional reads on newly released videos immediately without waiting for crowdsourced segment submissions.","tags":["creative-multimedia","domain-workflows"],"stars":81,"forks":12,"openIssues":0,"license":null,"language":"JavaScript","lastCommitAt":"2026-09-18T03:38:05Z","createdAt":"2026-09-17T23:46:01Z","metadataFetchedAt":"2026-09-21T22:07:42.068Z","avatarUrl":"https://avatars.githubusercontent.com/u/4214509?v=4","summarySource":"human-reviewed","claimStatus":"源码已核验。包含完整的 Chrome 扩展架构与 Jev 调用封装；无字幕视频依赖实时语音转录，耗费额外算力。","claimStatusEn":"Source code verified against pinned commit. Full extension source inspected; videos lacking transcripts require real-time client audio transcription.","plainSummaryJa":"YouTube 動画の字幕や音声を Jev で解析し、スポンサー広告セグメントを検出して自動スキップするブラウザー拡張機能です。","plainSummaryKo":"YouTube 영상의 자막과 오디오를 Jev로 분석하여 스폰서 광고 구간을 자동으로 감지하고 건너뛰는 브라우저 확장 프로그램입니다.","jevDecisionPointJa":"動画の字幕や音声テキストを Jev に送信して宣伝パートかどうかを判定し、プレイヤー側で再生タイムラインをスキップします。","jevDecisionPointKo":"자막 텍스트를 Jev에 전달하여 스폰서 광고 멘트 여부를 판정한 뒤 브라우저 코드가 타임라인을 자동으로 이동시킵니다.","highlightBenefitJa":"コミュニティによる手動登録を待たずに、新規動画でもリアルタイムに口頭広告を識別して快適に視聴できます。","highlightBenefitKo":"크라우드소싱 등록을 기다리지 않고도 새 영상의 구두 광고 구간을 즉시 감지하여 쾌적한 시청 환경을 제공합니다.","claimStatusJa":"固定コミットのソースを確認済み。拡張機能コードを確認。字幕が存在しない動画では音声書き起こしが必要となります。","claimStatusKo":"고정 커밋 기준 소스 확인 완료. 전체 확장 프로그램 소스를 확인했으나 자막이 없는 영상은 별도 음성 전사가 요구됩니다.","licenseStatus":"unconfirmed","sourceReviewedAt":"2026-09-19T13:00:00.000000+00:00","evidence":[{"url":"https://github.com/trungdq88/youtube-sponsor-detection/blob/de01f0568d043035889a296a61ce21e0accc8b16/extension/lib/jev.js#L1-L541","note":"extension/lib/jev.js (L1-L541)"},{"url":"https://github.com/trungdq88/youtube-sponsor-detection/blob/de01f0568d043035889a296a61ce21e0accc8b16/README.md","note":"README.md"}],"pinned":false},{"id":"jaredpalmer:kev","name":"kev","author":"jaredpalmer","url":"https://github.com/jaredpalmer/kev","category":"High-Frequency & Simulation","plainSummary":"基于 Qwen2.5-0.5B 构建的轻量级类 Jev 决策头与适配器，支持在 MacBook 本地训练、微调与端到端运行。","plainSummaryEn":"Tiny Jev-like decision head built atop Qwen2.5-0.5B that can be trained and run locally on Apple Silicon MacBooks.","jevDecisionPoint":"在小型基础模型上附加并联判断头，接收状态并直接输出强类型离散问题的概率分布，无需生成冗长文本。","jevDecisionPointEn":"Attaches a parallel decision head to an open 0.5B model to answer discrete questions directly from token activations.","highlightBenefit":"极度轻量化（仅 0.5B 参数），支持本地纯离线部署与低功耗边缘端决策，为探索专用判断模型提供了开源基座。","highlightBenefitEn":"Runs entirely offline on edge hardware with minimal memory footprint, offering an accessible base for custom judgment modeling.","tags":["evaluation-benchmarks","typed-decisions"],"stars":2347,"forks":121,"openIssues":2,"license":"Apache-2.0","language":"Python","lastCommitAt":"2026-09-21T22:00:29Z","createdAt":"2026-09-17T20:49:39Z","metadataFetchedAt":"2026-09-21T22:07:45.019Z","avatarUrl":"https://avatars.githubusercontent.com/u/4060187?v=4","summarySource":"human-reviewed","claimStatus":"源码已核验。包含模型权重训练代码与对比评估用例；模型容量较小，复杂语义多跳逻辑推理能力不及云端大模型。","claimStatusEn":"Source code verified against pinned commit. Complete training and evaluation pipeline; limited parameter capacity restricts multi-hop reasoning.","plainSummaryJa":"Qwen2.5-0.5B をベースに構築された軽量な Jev 風の判断ヘッドモデルであり、MacBook 上でローカル学習および推論が可能です。","plainSummaryKo":"Qwen2.5-0.5B 기반으로 구축된 초경량 Jev 스타일 판단 헤드 모델로 MacBook 환경에서 로컬 훈련과 추론이 가능합니다.","jevDecisionPointJa":"0.5B モデルの上に並列決定ヘッドを配置し、トークン活性から型付き離散質問の確率を直接算出します。","jevDecisionPointKo":"0.5B 오픈 모델 위에 병렬 결정 헤드를 추가하여 텍스트 생성 없이 토큰 활성화로부터 이산 질문의 확률을 산출합니다.","highlightBenefitJa":"完全なオフライン環境と省電力ハードウェアで動作し、独自データでの微調整やエッジ推論の基盤として役立ちます。","highlightBenefitKo":"에지 기기에서 완전 오프라인으로 구동되며 최소한의 메모리 점유율로 전용 판단 모델을 실험할 수 있는 기반을 제공합니다.","claimStatusJa":"固定コミットのソースを確認済み。学習パイプラインと評価コードを確認。パラメータ規模の制約上、複雑な推論には限界があります。","claimStatusKo":"고정 커밋 기준 소스 확인 완료. 전체 훈련 및 평가 파이프라인을 확인했으나 매개변수 규모로 인해 복합 추론에는 한계가 있습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T13:00:00.000000+00:00","evidence":[{"url":"https://github.com/jaredpalmer/kev/blob/cc954f2f66d86943688fdbfa07b6db18f82ed065/kev/jev.py#L1-L122","note":"kev/jev.py (L1-L122)"},{"url":"https://github.com/jaredpalmer/kev/blob/cc954f2f66d86943688fdbfa07b6db18f82ed065/README.md","note":"README.md"}],"pinned":false},{"id":"tianyucodings:nanojev","name":"NanoJev","author":"TianyuCodings","url":"https://github.com/TianyuCodings/NanoJev","category":"High-Frequency & Simulation","plainSummary":"Jev 决策模型的纳米级复刻版本，支持并行决策输出、动态候选集与完整的端到端训练评估流水线。","plainSummaryEn":"Nano replica of Jev featuring parallel decisions, dynamic candidate sets, and an end-to-end training and evaluation pipeline.","jevDecisionPoint":"通过前向网络对多个并行问题与动态候选集进行一次性评分，输出归一化置信度并记录游戏环境中的动作路径。","jevDecisionPointEn":"Evaluates multiple questions and dynamic candidate spaces concurrently in a single forward pass, logging navigation choices.","highlightBenefit":"开源了从数据准备、模型训练到游戏评测的全套流程，为学术界与开源社区研究非生成式离散决策模型提供参考。","highlightBenefitEn":"Provides open-source training infrastructure and game environment benchmarks for investigating non-generative discrete decisions.","tags":["games-simulation","evaluation-benchmarks","typed-decisions"],"stars":1795,"forks":198,"openIssues":7,"license":"MIT","language":"Python","lastCommitAt":"2026-09-21T09:58:22Z","createdAt":"2026-09-17T16:08:30Z","metadataFetchedAt":"2026-09-21T22:07:48.011Z","avatarUrl":"https://avatars.githubusercontent.com/u/48460209?v=4","summarySource":"human-reviewed","claimStatus":"源码已核验。包含训练脚本与模拟导航评估代码；复刻模型的训练收敛度与各下游任务上的泛化度需独立验证。","claimStatusEn":"Source code verified against pinned commit. Complete training and benchmark suite; generalization depends on dataset diversity.","plainSummaryJa":"並列決定出力、動的候補セット、エンドツーエンドの学習・評価パイプラインを備えた Jev のナノスケール再現プロジェクトです。","plainSummaryKo":"병렬 결정 출력, 동적 후보군 처리 및 종단간 훈련과 평가 파이프라인을 갖춘 Jev 나노 재현 프로젝트입니다.","jevDecisionPointJa":"1 回のフォワードパスで複数の並列質問と動的選択肢を同時に評価し、環境内でのナビゲーション判断を記録します。","jevDecisionPointKo":"단일 순전파 연산으로 다중 병렬 질문과 동적 선택지를 동시에 평가하여 내비게이션 결정을 도출합니다.","highlightBenefitJa":"非生成型の離散判断モデルに関するデータセット作成から学習、ゲーム環境での評価までの全工程をオープンに提供します。","highlightBenefitKo":"텍스트를 생성하지 않는 이산 판단 모델에 대한 데이터셋 구축, 훈련, 게임 환경 평가 전 과정을 공개된 코드로 제공합니다.","claimStatusJa":"固定コミットのソースを確認済み。学習スクリプトと評価環境を確認。下流タスクでの汎化性能は学習データに依存します。","claimStatusKo":"고정 커밋 기준 소스 확인 완료. 훈련 및 평가 코드를 확인했으나 하위 과제에서의 일반화 성능은 데이터 다양성에 좌우됩니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T13:00:00.000000+00:00","evidence":[{"url":"https://github.com/TianyuCodings/NanoJev/blob/71a513bb0163b5634467842b523ee0c0ed6fb1c7/scripts/evaluate_live_jev_navigation.py#L1-L70","note":"scripts/evaluate_live_jev_navigation.py (L1-L70)"},{"url":"https://github.com/TianyuCodings/NanoJev/blob/71a513bb0163b5634467842b523ee0c0ed6fb1c7/README.zh-CN.md","note":"README.zh-CN.md (L1-L186)"},{"url":"https://github.com/TianyuCodings/NanoJev/blob/71a513bb0163b5634467842b523ee0c0ed6fb1c7/README.md","note":"README.md"}],"pinned":false},{"id":"featherless-ai:simple-jev","name":"simple-jev","author":"featherless-ai","url":"https://github.com/featherless-ai/simple-jev","category":"SDK & Decision Frameworks","plainSummary":"将任意开源大语言模型转化为分类器与 Jev 兼容端点的适配服务，无需额外训练专用分类头。","plainSummaryEn":"Adapter turning open LLM endpoints into Jev-compatible classification services without training a separate classifier head.","jevDecisionPoint":"通过分析输入文本在模型词表中的 logits 分布，将候选项映射为标准 Jev 离散概率输出。","jevDecisionPointEn":"Extracts log-probabilities of candidate tokens from model vocabulary logits, formatting them into standard Jev responses.","highlightBenefit":"允许开发者使用现有的 Hugging Face 或 vLLM 兼容服务快速搭建私有 Jev 决策代理，降低尝试门槛。","highlightBenefitEn":"Enables rapid prototyping of Jev-like decision workflows against standard inference servers such as vLLM or Hugging Face.","tags":["multilanguage-sdk","typed-decisions"],"stars":447,"forks":49,"openIssues":1,"license":"Apache-2.0","language":"Python","lastCommitAt":"2026-09-21T04:41:35Z","createdAt":"2026-09-18T09:52:15Z","metadataFetchedAt":"2026-09-21T22:07:51.058Z","avatarUrl":"https://avatars.githubusercontent.com/u/172633431?v=4","summarySource":"human-reviewed","claimStatus":"源码已核验。包含 API 适配层与 jevpilot 演示代码；对分词器敏感度较高，且模型内在偏置会影响概率校准度。","claimStatusEn":"Source code verified against pinned commit. API adapter and demo scripts inspected; output probability calibration depends on base model.","plainSummaryJa":"独自の分類ヘッドを学習させることなく、オープン LLM を Jev 互換の分類エンドポイントに変換するアダプターです。","plainSummaryKo":"별도의 분류 헤드 학습 없이 오픈 언어 모델을 Jev 호환 분류 엔드포인트로 변환해 주는 어댑터 서비스입니다.","jevDecisionPointJa":"モデルのボキャブラリ logits から候補トークンの対数確率を抽出し、標準的な Jev 決定レスポンスとして成形します。","jevDecisionPointKo":"모델의 어휘 로짓에서 후보 토큰의 로그 확률을 추출하여 표준 Jev 판단 응답 규격으로 변환합니다.","highlightBenefitJa":"既存の vLLM や Hugging Face 推論基盤を活用して、迅速にプライベートな Jev 決定プロトタイプを構築できます。","highlightBenefitKo":"기존 vLLM이나 Hugging Face 추론 인프라를 활용하여 사설 Jev 스타일 판단 서비스를 손쉽게 구축할 수 있습니다.","claimStatusJa":"固定コミットのソースを確認済み。アダプター実装とデモコードを確認。確率の較正精度は基盤モデルのバイアスに依存します。","claimStatusKo":"고정 커밋 기준 소스 확인 완료. 어댑터 및 데모 코드를 확인했으나 확률 보정 품질은 기본 모델의 특성에 영향을 받습니다.","licenseStatus":"unconfirmed","sourceReviewedAt":"2026-09-19T13:00:00.000000+00:00","evidence":[{"url":"https://github.com/featherless-ai/simple-jev/blob/0dd5396ffce671ab7c4bfc031506d8e558cf8d23/demos/jevpilot/src/simple-jev-api.js#L1-L85","note":"demos/jevpilot/src/simple-jev-api.js (L1-L85)"},{"url":"https://github.com/featherless-ai/simple-jev/blob/0dd5396ffce671ab7c4bfc031506d8e558cf8d23/demos/jevpilot/src/jev-request.js","note":"demos/jevpilot/src/jev-request.js (L1-L398)"},{"url":"https://github.com/featherless-ai/simple-jev/blob/0dd5396ffce671ab7c4bfc031506d8e558cf8d23/README.md","note":"README.md"}],"pinned":false},{"id":"firassx914:janus","name":"Janus","author":"FirasSX914","url":"https://github.com/FirasSX914/Janus","category":"Routing & Cost Optimization","plainSummary":"模型适用性测量与路由框架，评估业务数据在 Jev 与传统大模型之间的收益边界并执行最优动态分发。","plainSummaryEn":"Framework for measuring when to employ Jev versus generative LLMs on proprietary datasets, routing queries based on measured benchmarks.","jevDecisionPoint":"通过内置的 TypeSafe 提供者向 Jev 发起基准测试请求，统计离散判断准确度与延迟，生成自动化路由策略。","jevDecisionPointEn":"Benchmarks Jev response accuracy and latency via a TypeSafe provider integration, deriving empirical routing policies.","highlightBenefit":"用实测数据消除模型选型主观臆断，在保证任务成功率的同时实现系统整体调用的成本最小化。","highlightBenefitEn":"Replaces intuition with empirical data when adopting discrete decision models, optimizing overall cost while preserving task success.","tags":["llm-routing-cost","evaluation-benchmarks"],"stars":2,"forks":0,"openIssues":1,"license":"MIT","language":"Python","lastCommitAt":"2026-09-18T13:54:36Z","createdAt":"2026-09-17T12:44:54Z","metadataFetchedAt":"2026-09-21T22:07:54.086Z","avatarUrl":"https://avatars.githubusercontent.com/u/66559814?v=4","summarySource":"human-reviewed","claimStatus":"源码已核验。包含 TypeSafe 提供者实现与路由策略引擎；实际成本节省幅度取决于业务查询的数据特征。","claimStatusEn":"Source code verified against pinned commit. Provider code and evaluation policies verified; actual savings depend on dataset profiles.","plainSummaryJa":"自社データ上で Jev と通常の大規模言語モデルの適性を計測し、最適な振り分けルールを導出して実行するルーティングフレームワークです。","plainSummaryKo":"자체 데이터셋에서 Jev와 일반 생성형 모델의 효율을 측정하고 실측치에 기반하여 최적 경로로 분기하는 라우팅 프레임워크입니다.","jevDecisionPointJa":"TypeSafe プロバイダー経由で Jev の精度とレイテンシをベンチマーク測定し、経験的な自動ルーティングルールを生成します。","jevDecisionPointKo":"TypeSafe 공급자 연동을 통해 Jev의 응답 정확도와 지연 시간을 벤치마킹하여 경험적 분기 정책을 도출합니다.","highlightBenefitJa":"経験則ではなく実測データに基づいて Jev の導入箇所を特定し、タスク成功率を維持しながらシステム全体のコストを最小化します。","highlightBenefitKo":"추측 대신 실측 데이터에 기반하여 판단 모델을 적용할 영역을 규명하고 비용 대비 성능을 최적화합니다.","claimStatusJa":"固定コミットのソースを確認済み。プロバイダー実装とルーティング基盤を確認。コスト削減幅はデータの特性に依存します。","claimStatusKo":"고정 커밋 기준 소스 확인 완료. 공급자 연동 및 정책 엔진을 확인했으나 실제 비용 절감 폭은 데이터 특성에 따라 달라집니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T13:00:00.000000+00:00","evidence":[{"url":"https://github.com/FirasSX914/Janus/blob/9cb66c488cf884e5997356a0de45f75aa3148774/src/janus/providers/typesafe.py#L1-L60","note":"src/janus/providers/typesafe.py (L1-L60)"},{"url":"https://github.com/FirasSX914/Janus/blob/9cb66c488cf884e5997356a0de45f75aa3148774/README.md","note":"README.md"}],"pinned":false,"topics":["benchmark","calibration","confidence","deepseek","jev","llm","routing","typesafe"]},{"id":"vinilana:jev-eval-agent","name":"jev-eval-agent","author":"vinilana","url":"https://github.com/vinilana/jev-eval-agent","category":"Domain & Vertical Tools","plainSummary":"智能体工具选择基准测试平台，在包含 100 个模拟工具的个人助理环境下对比大模型直接选工具与 Jev 路由的效率。","plainSummaryEn":"Agent tool evaluation harness comparing standard LLM tool selection against Jev routing across 100 mocked tools.","jevDecisionPoint":"利用 Jev 对上百个工具元数据进行两级离散筛选和相关性判定，把候选工具集从百级快速裁剪至少量候选。","jevDecisionPointEn":"Employs Jev for two-stage discrete pruning over tool catalogs, narrowing extensive tool lists down to relevant candidates in milliseconds.","highlightBenefit":"解决了多工具智能体因长提示词导致的上下文膨胀与选择幻觉问题，显著降低了大模型的规划耗时。","highlightBenefitEn":"Mitigates context bloat and hallucinated tool calls in complex agents by isolating candidate selection into fast discrete steps.","tags":["evaluation-benchmarks","coding-agents"],"stars":102,"forks":10,"openIssues":1,"license":null,"language":"HTML","lastCommitAt":"2026-09-17T11:07:29Z","createdAt":"2026-09-17T05:13:21Z","metadataFetchedAt":"2026-09-21T22:07:57.033Z","avatarUrl":"https://avatars.githubusercontent.com/u/17963870?v=4","summarySource":"human-reviewed","claimStatus":"源码已核验。包含完整的 Agent 路由实现与模拟测试环境；测试工具目前主要为 mock 状态，非生产级全真工具链。","claimStatusEn":"Source code verified against pinned commit. Agent routing and simulated suite inspected; tools are primarily mocked environments.","plainSummaryJa":"100 種類のモックツールを備えたエージェント環境において、通常の LLM によるツール選択と Jev ルーティングの性能を比較評価する検証基盤です。","plainSummaryKo":"100개 모의 도구를 갖춘 에이전트 환경에서 기존 대규모 모델의 도구 선택과 Jev 라우팅의 효율을 비교 평가하는 벤치마크 도구입니다.","jevDecisionPointJa":"Jev を用いて大量のツール候補から 2 段階の離散選択と関連性判定を行い、候補ツールを高速に絞り込みます。","jevDecisionPointKo":"Jev를 활용하여 방대한 도구 목록에서 2단계 이산 필터링을 거쳐 유효 후보 도구를 밀리초 단위로 압축합니다.","highlightBenefitJa":"ツール数の増加に伴うプロンプト肥大化やツールの誤認呼び出しを抑制し、エージェントの推論効率を向上させます。","highlightBenefitKo":"도구 증가에 따른 프롬프트 비대화와 잘못된 도구 호출 오류를 방지하고 에이전트의 계획 수립 속도를 높입니다.","claimStatusJa":"固定コミットのソースを確認済み。エージェントルーターとシミュレーション環境を確認。ツールは主にモック実装です。","claimStatusKo":"고정 커밋 기준 소스 확인 완료. 에이전트 라우팅 및 시뮬레이션 환경을 확인했으나 실제 도구는 모의 환경 위주입니다.","licenseStatus":"unconfirmed","sourceReviewedAt":"2026-09-19T13:00:00.000000+00:00","evidence":[{"url":"https://github.com/vinilana/jev-eval-agent/blob/037de1120c84b4b63cdf748e2acf258ff66d7731/agent/lib/jev-router.ts#L1-L154","note":"agent/lib/jev-router.ts (L1-L154)"},{"url":"https://github.com/vinilana/jev-eval-agent/blob/037de1120c84b4b63cdf748e2acf258ff66d7731/README.md","note":"README.md"}],"pinned":false},{"id":"zhuyansen:jev-search-rerank-eval","name":"jev-search-rerank-eval","author":"zhuyansen","url":"https://github.com/zhuyansen/jev-search-rerank-eval","category":"Data & Search","plainSummary":"中英文检索重排效果评估系统，对比 Jev 重排与词法搜索、向量检索及混合融合基线，并分析评审者自循环偏差。","plainSummaryEn":"Retrieval evaluation system measuring Jev reranking against lexical, embedding, and fusion baselines across 9,831 query-document pairs.","jevDecisionPoint":"在评估流水线中调用 Jev 判断器对 9,831 对样本和 164 个中英查询进行多级相关度评分，比较重排前后排序指标。","jevDecisionPointEn":"Invokes Jev judges to grade multi-level relevance on bilingual queries, evaluating ranking metrics and judge-circularity effects.","highlightBenefit":"为搜索算法工程师提供了严谨的定性与定量对比依据，以实测实验揭示了 Jev 在跨语种技能目录搜索中的真实收益与边界。","highlightBenefitEn":"Provides rigorous empirical evidence comparing discrete judgment rerankers against vector search in multilingual discovery contexts.","tags":["search-retrieval","evaluation-benchmarks","classification-ranking"],"stars":6,"forks":0,"openIssues":1,"license":"MIT","language":"Python","lastCommitAt":"2026-09-18T08:59:59Z","createdAt":"2026-09-18T09:00:07Z","metadataFetchedAt":"2026-09-21T22:08:00.038Z","avatarUrl":"https://avatars.githubusercontent.com/u/24306647?v=4","summarySource":"human-reviewed","claimStatus":"源码已核验。包含完整的评估套件、基准数据提取与统计脚本；评测数据集来自 Agent Skills Hub，不同领域需重新校准。","claimStatusEn":"Source code verified against pinned commit. Complete evaluation suite and CLI inspected; benchmarked on Agent Skills Hub data.","plainSummaryJa":"9,831 件のペアと 164 件の日英クエリを用いて、Jev リランクと単語検索、ベクトル検索、ハイブリッド検索を比較検証する評価システムです。","plainSummaryKo":"9,831개 쌍과 164개 한/영 쿼리를 바탕으로 Jev 리랭킹과 어휘 검색, 벡터 검색, 하이브리드 검색을 다각도로 비교 분석하는 평가 시스템입니다.","jevDecisionPointJa":"評価パイプライン内で Jev 判定器を呼び出し、複数段階の関連度スコアを算出して順位指標や判定者の循環バイアスを測定します。","jevDecisionPointKo":"평가 파이프라인에서 Jev 판정기를 호출하여 다단계 관련도 점수를 부여하고 순위 지표와 심사 순환 편향을 측정합니다.","highlightBenefitJa":"多言語カタログ検索において、離散判断リランカーがベクトル検索に対してどの程度優位性を持つかを実証的に明らかにします。","highlightBenefitKo":"다국어 카탈로그 검색에서 이산 판단 리랭커가 임베딩 검색 대비 갖는 실제 개선 효과와 한계를 실증 데이터로 검증합니다.","claimStatusJa":"固定コミットのソースを確認済み。評価スイートと CLI スクリプトを確認。データセットは Agent Skills Hub に準拠しています。","claimStatusKo":"고정 커밋 기준 소스 확인 완료. 전체 평가 도구와 CLI를 확인했으나 데이터셋은 Agent Skills Hub 기준에 맞춰져 있습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T13:00:00.000000+00:00","evidence":[{"url":"https://github.com/zhuyansen/jev-search-rerank-eval/blob/c896f14e944182e17a002dd7546f06ba3788d586/src/jse/rankers/jev.py#L1-L46","note":"src/jse/rankers/jev.py (L1-L46)"},{"url":"https://github.com/zhuyansen/jev-search-rerank-eval/blob/c896f14e944182e17a002dd7546f06ba3788d586/src/jse/label/jev_judge.py","note":"src/jse/label/jev_judge.py (L1-L19)"},{"url":"https://github.com/zhuyansen/jev-search-rerank-eval/blob/c896f14e944182e17a002dd7546f06ba3788d586/src/jse/__init__.py","note":"src/jse/__init__.py (L1-L30)"},{"url":"https://github.com/zhuyansen/jev-search-rerank-eval/blob/c896f14e944182e17a002dd7546f06ba3788d586/src/jse/__main__.py","note":"src/jse/__main__.py (L1-L4)"},{"url":"https://github.com/zhuyansen/jev-search-rerank-eval/blob/c896f14e944182e17a002dd7546f06ba3788d586/src/jse/catalog.py","note":"src/jse/catalog.py (L1-L64)"},{"url":"https://github.com/zhuyansen/jev-search-rerank-eval/blob/c896f14e944182e17a002dd7546f06ba3788d586/src/jse/cli.py","note":"src/jse/cli.py (L1-L173)"},{"url":"https://github.com/zhuyansen/jev-search-rerank-eval/blob/c896f14e944182e17a002dd7546f06ba3788d586/src/jse/evaluate.py","note":"src/jse/evaluate.py (L1-L146)"},{"url":"https://github.com/zhuyansen/jev-search-rerank-eval/blob/c896f14e944182e17a002dd7546f06ba3788d586/README.md","note":"README.md"}],"pinned":false,"topics":["bge-m3","evaluation","information-retrieval","ndcg","reranking","typesafe-jev"]},{"id":"pnthn-ai:polar_llama","name":"polar_llama","author":"pnthn-ai","url":"https://github.com/pnthn-ai/polar_llama","category":"Data & Search","plainSummary":"Polars 上的并行推理库：聊天模型走各家补全接口，Jev 则按行做 Noul、Choice、Score，或对整份文档套一份类型化契约。","plainSummaryEn":"A Polars library for parallel provider inference that also calls Jev per row as Noul, Choice, and Score questions, or as one typed contract over a document.","jevDecisionPoint":"每一行一个 state，多个 typed questions 一次请求返回；Noul / Choice / Score 落成带置信度的普通列。","jevDecisionPointEn":"One System One request per row: Noul, Choice, or Score answers land as ordinary typed columns with confidences.","highlightBenefit":"把闭集判断接到现有 Polars 批处理列上，而不必再走一轮聊天补全。","highlightBenefitEn":"Puts closed-set Jev answers on existing Polars columns instead of another chat-completion pass.","tags":["search-retrieval","classification-ranking"],"stars":30,"forks":2,"openIssues":10,"license":"MIT","lastCommitAt":"2026-09-19T17:21:10Z","createdAt":"2024-05-03T16:47:23Z","metadataFetchedAt":"2026-09-21T22:08:03.257Z","avatarUrl":"https://avatars.githubusercontent.com/u/295237610?v=4","summarySource":"curated","claimStatus":"优先保留投稿者与仓库原文，缺失语言自动补充；自动检查仅确认 Jev 集成证据，未经本站运行或性能复测。","claimStatusEn":"Author and repository text is preserved where clear; missing languages are enriched automatically. Checks establish source-level Jev integration, not runtime, safety, or performance validation.","plainSummaryJa":"Polars 上の並列推論ライブラリ。チャットモデルは各社の補完 API を使い、Jev は行ごとに Noul、Choice、Score、または文書全体に型付き契約を当てる。","plainSummaryKo":"Polars 위의 병렬 추론 라이브러리. 채팅 모델은 각 제공자 완성 API를 쓰고, Jev는 행마다 Noul, Choice, Score를 하거나 문서 전체에 타입이 있는 계약을 씌운다.","jevDecisionPointJa":"行ごとに 1 つの state。複数の typed questions を 1 リクエストで返し、Noul / Choice / Score を信頼度付きの普通の列にする。","jevDecisionPointKo":"행마다 state 하나. typed questions 여러 개를 한 요청으로 돌려 Noul / Choice / Score를 신뢰도가 있는 일반 열로 만든다.","highlightBenefitJa":"閉集合の判断を既存の Polars バッチ列に載せ、チャット補完をもう一度走らせなくてよい。","highlightBenefitKo":"닫힌 집합 판단을 기존 Polars 배치 열에 붙여, 채팅 완성을 한 번 더 돌리지 않아도 된다.","evidence":[{"url":"https://github.com/pnthn-ai/polar_llama/blob/d10d48db20af9d5a33e696c39f660187677257f7/README.md","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/pnthn-ai/polar_llama/blob/d10d48db20af9d5a33e696c39f660187677257f7/src/model_client/typesafe.rs","note":"固定版本的 Jev 集成与说明来源"}]},{"id":"chy4pro:jevbrowserext","name":"JevBrowserExt","author":"chy4pro","url":"https://github.com/chy4pro/JevBrowserExt","category":"Browser & OS Action","plainSummary":"把 jev-ultrafast 做成 Manifest V3 Chrome 扩展：Jev 在当前标签里选操作和 DOM 元素，只有输入文字时才调用小型对话模型。","plainSummaryEn":"A Manifest V3 Chrome port of jev-ultrafast: Jev picks the operation and DOM element in one request per step; a small chat model fills TYPE_TEXT.","jevDecisionPoint":"一次请求选择 CLICK、TYPE_TEXT、SELECT、SCROLL_DOWN、SCROLL_UP、PRESS_ENTER、WAIT、DONE 或 BLOCKED 以及对应元素；PRESS_ENTER 是独立按键控件。另用独立是非题核对目标是否已达成、动作是否卡住。","jevDecisionPointEn":"Chooses CLICK, TYPE_TEXT, SELECT, SCROLL_DOWN, SCROLL_UP, PRESS_ENTER, WAIT, DONE, or BLOCKED and the matching DOM target in one request; PRESS_ENTER is a separate key control. Separate yes/no checks cover goal completion and a stuck loop.","highlightBenefit":"在用户自己的标签里跑，不截图；操作、目标元素和输入文案可以分开检查。","highlightBenefitEn":"Runs in the user’s own tabs without screenshots, and keeps the action, target, and typed string inspectable.","tags":["browser-automation","typed-decisions"],"stars":15,"forks":1,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-21T08:09:29Z","createdAt":"2026-09-18T21:24:11Z","metadataFetchedAt":"2026-09-21T22:08:09.034Z","avatarUrl":"https://avatars.githubusercontent.com/u/61305850?v=4","summarySource":"curated","claimStatus":"优先保留投稿者与仓库原文，缺失语言自动补充；自动检查仅确认 Jev 集成证据，未经本站运行或性能复测。","claimStatusEn":"Author and repository text is preserved where clear; missing languages are enriched automatically. Checks establish source-level Jev integration, not runtime, safety, or performance validation.","plainSummaryJa":"jev-ultrafast を Manifest V3 Chrome 拡張にしたもの。Jev が今のタブで操作と DOM 要素を選び、文字入力が必要なときだけ小型の対話モデルを呼ぶ。","plainSummaryKo":"jev-ultrafast를 Manifest V3 Chrome 확장으로 옮긴 것. Jev가 현재 탭에서 동작과 DOM 요소를 고르고, 글자를 넣을 때만 작은 대화 모델을 부른다.","jevDecisionPointJa":"1 リクエストで CLICK、TYPE_TEXT、SELECT、SCROLL_DOWN、SCROLL_UP、PRESS_ENTER、WAIT、DONE、BLOCKED と対応要素を選ぶ。PRESS_ENTER は独立したキー操作。目標達成と動作停滞は別の是非問で確認する。","jevDecisionPointKo":"한 요청에서 CLICK, TYPE_TEXT, SELECT, SCROLL_DOWN, SCROLL_UP, PRESS_ENTER, WAIT, DONE, BLOCKED와 해당 요소를 고른다. PRESS_ENTER는 별도의 키 컨트롤이다. 목표 달성과 동작 정체는 별도의 예/아니오로 확인한다.","highlightBenefitJa":"ユーザー自身のタブで動き、スクリーンショットは撮らない。操作、対象要素、入力文を分けて確認できる。","highlightBenefitKo":"사용자 자신의 탭에서 돌고 스크린샷은 찍지 않는다. 동작, 대상 요소, 입력 문장을 나눠 볼 수 있다.","evidence":[{"url":"https://github.com/chy4pro/JevBrowserExt/blob/0e4c5294ab3de8db09e6ddbc523ce8004406e199/README.md","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/chy4pro/JevBrowserExt/blob/0e4c5294ab3de8db09e6ddbc523ce8004406e199/README_CN.md","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/chy4pro/JevBrowserExt/blob/0e4c5294ab3de8db09e6ddbc523ce8004406e199/src/shared/providers/typesafe.ts","note":"固定版本的 Jev 集成与说明来源"}]},{"id":"newuser7171:antivirus","name":"antivirus","author":"newuser7171","url":"https://github.com/newuser7171/antivirus","category":"Security & Guardrails","plainSummary":"从文件静态特征构造状态，让 Jev 给出裁决、0–4 严重度和若干是非指标，再由本地规则决定隔离、放行或复核。","plainSummaryEn":"A file scanner that sends extracted features to Jev for a verdict, a 0–4 severity score, and Noul indicators, then applies local quarantine or review rules.","jevDecisionPoint":"Choice：clean / suspicious_pua / malicious；Score：威胁 0–4；Noul：加壳混淆、远控下载、持久化、注入规避。","jevDecisionPointEn":"Choice clean / suspicious_pua / malicious, Score 0–4, and Noul questions for packing, C2, persistence, and injection; local policy maps the answers.","highlightBenefit":"把启发式特征变成可设阈值的结构化判断，而不是一段杀毒评语；本站未运行或验证防护效果。","highlightBenefitEn":"Turns heuristic file features into thresholded structured answers rather than a prose verdict; this site has not run or validated protection.","tags":["security-guardrails","classification-ranking"],"stars":2,"forks":0,"openIssues":1,"license":null,"lastCommitAt":"2026-09-20T16:42:13Z","createdAt":"2026-09-18T14:48:02Z","metadataFetchedAt":"2026-09-21T22:08:12.073Z","avatarUrl":"https://avatars.githubusercontent.com/u/217369131?v=4","summarySource":"curated","claimStatus":"优先保留投稿者与仓库原文，缺失语言自动补充；自动检查仅确认 Jev 集成证据，未经本站运行或性能复测。","claimStatusEn":"Author and repository text is preserved where clear; missing languages are enriched automatically. Checks establish source-level Jev integration, not runtime, safety, or performance validation.","plainSummaryJa":"ファイルの静的特徴から state を作り、Jev に判定、0–4 の重大度、いくつかの是非指標を出させ、ローカル規則が隔離・通過・再確認を決める。","plainSummaryKo":"파일 정적 특징으로 state를 만들어 Jev에게 판정, 0–4 심각도, 몇 가지 예/아니오 지표를 받게 한 뒤, 로컬 규칙이 격리, 통과, 재확인을 정한다.","jevDecisionPointJa":"Choice は clean / suspicious_pua / malicious。Score は脅威 0–4。Noul は梱包難読化、遠隔ダウンロード、永続化、注入回避。","jevDecisionPointKo":"Choice는 clean / suspicious_pua / malicious. Score는 위협 0–4. Noul은 패킹 난독화, 원격 다운로드, 지속성, 주입 회피.","highlightBenefitJa":"ヒューリスティックな特徴をしきい値付きの構造化判断にし、アンチウイルスの評論文にはしない。本サイトは実行も防護効果の検証もしていない。","highlightBenefitKo":"휴리스틱 특징을 임계값 있는 구조화 판단으로 만들고 백신 평문으로 쓰지 않는다. 이 사이트는 실행하거나 방어 효과를 검증하지 않았다.","licenseStatus":"unconfirmed","evidence":[{"url":"https://github.com/newuser7171/antivirus/blob/4a9ea18ecf6c5c77991aa584aa86eb2822e8b6c9/README.md","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/newuser7171/antivirus/blob/4a9ea18ecf6c5c77991aa584aa86eb2822e8b6c9/jev_scanner.py","note":"固定版本的 Jev 集成与说明来源"}]},{"id":"coo-quack:jev-pii-checker","name":"jev-pii-checker","author":"coo-quack","url":"https://github.com/coo-quack/jev-pii-checker","category":"Security & Guardrails","plainSummary":"把文本交给 TypeSafe Jev 做 PII 类别 Noul 和敏感度 Score，再用正则与分词标出跨度。","plainSummaryEn":"A CLI that sends text to TypeSafe Jev for PII category Nouls and a sensitivity Score, then locates spans with regex and segmentation.","jevDecisionPoint":"对每个文本块并行询问 12 类 PII 是否出现，并给出 none / low / high 敏感度。","jevDecisionPointEn":"For each text chunk, asks twelve PII-category Nouls and a three-level sensitivity Score.","highlightBenefit":"判定在 TypeSafe 服务器上完成；本地只提取跨度，不能当作离线保密扫描。","highlightBenefitEn":"Judgment runs on TypeSafe servers; local code only locates spans, so this is not an offline privacy scanner.","tags":["security-guardrails","typed-decisions","cli-git-gates"],"stars":2,"forks":0,"openIssues":1,"license":"MIT","language":"TypeScript","lastCommitAt":"2026-09-20T05:12:06Z","createdAt":"2026-09-19T13:57:58Z","metadataFetchedAt":"2026-09-21T22:08:15.058Z","avatarUrl":"https://avatars.githubusercontent.com/u/260753136?v=4","summarySource":"human-reviewed","claimStatus":"固定版本的公开源码已核对。本站未执行该项目，也未独立测量性能或安全性。","claimStatusEn":"Pinned public source was reviewed. This site did not execute the project or independently measure performance or safety.","plainSummaryJa":"テキストを TypeSafe Jev に送り、PII 種別の Noul と感度 Score を取り、正規表現と分かち書きで範囲を付ける CLI。","plainSummaryKo":"텍스트를 TypeSafe Jev에 보내 PII 종류 Noul과 민감도 Score를 받은 뒤, 정규식과 분절로 구간을 표시하는 CLI.","jevDecisionPointJa":"テキスト塊ごとに 12 種類の PII Noul と none/low/high の感度 Score を同時に尋ねる。","jevDecisionPointKo":"텍스트 조각마다 12가지 PII Noul과 none/low/high 민감도 Score를 함께 묻는다.","highlightBenefitJa":"判定は TypeSafe 側で行われ、ローカルは範囲抽出のみ。オフラインの機密スキャンではない。","highlightBenefitKo":"판정은 TypeSafe 서버에서 이뤄지고 로컬은 구간 추출만 한다. 오프라인 기밀 스캐너가 아니다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T15:50:00.000Z","evidence":[{"url":"https://github.com/coo-quack/jev-pii-checker/blob/6ad02cdc0d7ca563897ce9eff881b7a7b017d426/src/judge.ts","note":"固定版本源码已核对；本站未执行该项目。"},{"url":"https://github.com/coo-quack/jev-pii-checker/blob/6ad02cdc0d7ca563897ce9eff881b7a7b017d426/src/gate.ts","note":"固定版本源码已核对；本站未执行该项目。"},{"url":"https://github.com/coo-quack/jev-pii-checker/blob/6ad02cdc0d7ca563897ce9eff881b7a7b017d426/README.md","note":"固定版本源码已核对；本站未执行该项目。"},{"url":"https://x.com/chata/status/2101331708662698304","note":"X discovery lead; integration is proven by the pinned GitHub files, not by the post."}],"pinned":false,"topics":["ai-tools","bun","cli","jev","pii","privacy","typesafe","typescript"]},{"id":"fazalaali:jev-robotics-demo","name":"jev-robotics-demo","author":"FazalAAli","url":"https://github.com/FazalAAli/jev-robotics-demo","category":"High-Frequency & Simulation","plainSummary":"MuJoCo 机械臂叠方块演示：程序提出候选动作，Jev 选择目标、抓放和是否完成。","plainSummaryEn":"A MuJoCo arm demo where local code proposes candidate moves and Jev chooses the target, grasp or release, and completion.","jevDecisionPoint":"在候选目标、是否抓放、任务是否完成之间做 Choice 和 Noul。","jevDecisionPointEn":"Chooses among candidate targets, grasp or release, and a done Noul.","highlightBenefit":"动作候选在本地物理副本里生成，选择交给 Jev；作者对照数字未经本站复测。","highlightBenefitEn":"Candidate motions are generated in a local physics copy; Jev only chooses. Author-reported timings were not retested here.","tags":["games-simulation","typed-decisions","domain-workflows"],"stars":3,"forks":0,"openIssues":0,"license":"MIT","language":"Python","lastCommitAt":"2026-09-19T15:08:10Z","createdAt":"2026-09-19T15:09:19Z","metadataFetchedAt":"2026-09-21T22:08:18.040Z","avatarUrl":"https://avatars.githubusercontent.com/u/37760883?v=4","summarySource":"human-reviewed","claimStatus":"固定版本的公开源码已核对。本站未执行该项目，也未独立测量性能或安全性。","claimStatusEn":"Pinned public source was reviewed. This site did not execute the project or independently measure performance or safety.","plainSummaryJa":"MuJoCo のアームが積み木をするデモ。プログラムが候補動作を出し、Jev が目標・把持/解放・完了を選ぶ。","plainSummaryKo":"MuJoCo 팔 시연: 프로그램이 후보 동작을 내고 Jev가 목표, 집기/놓기, 완료 여부를 고른다.","jevDecisionPointJa":"候補目標、把持/解放、完了の Noul を Choice / Noul で選ぶ。","jevDecisionPointKo":"후보 목표, 집기/놓기, 완료 Noul을 Choice와 Noul로 고른다.","highlightBenefitJa":"候補動作はローカル物理コピーで作り、選択だけ Jev に渡す。作者の所要時間は未再測。","highlightBenefitKo":"후보 동작은 로컬 물리 사본에서 만들고 선택만 Jev에 맡긴다. 저자 시간 수치는 여기서 재측정하지 않았다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T15:50:00.000Z","evidence":[{"url":"https://github.com/FazalAAli/jev-robotics-demo/blob/531de61a75f386b0847f5f6f809a11424a75c29b/jev_agent.py","note":"固定版本源码已核对；本站未执行该项目。"},{"url":"https://github.com/FazalAAli/jev-robotics-demo/blob/531de61a75f386b0847f5f6f809a11424a75c29b/README.md","note":"固定版本源码已核对；本站未执行该项目。"},{"url":"https://x.com/FazAliDev/status/2101327974335422525","note":"X discovery lead; integration is proven by the pinned GitHub files, not by the post."}],"pinned":false},{"id":"uehaj:jev-semgrep","name":"jev-semgrep","author":"uehaj","url":"https://github.com/uehaj/jev-semgrep","category":"Data & Search","plainSummary":"用 Jev 给每一行打“是否符合某含义”的分，可用 AND/OR/NOT 组合，并支持跨语言查询。","plainSummaryEn":"A semantic grep that scores each line with Jev against a meaning, including AND/OR/NOT and cross-language queries.","jevDecisionPoint":"每批约 30 行，对每行向 `jev-latest` 发 Score 或 Noul，问它是否匹配给定含义。","jevDecisionPointEn":"Batches about thirty lines and asks jev-latest a Score or Noul per line for the stated meaning.","highlightBenefit":"零运行时依赖的语义 grep；查询文本会发到 TypeSafe。","highlightBenefitEn":"A zero-runtime-dependency semantic grep; query text is sent to TypeSafe.","tags":["search-retrieval","cli-git-gates","typed-decisions"],"stars":123,"forks":7,"openIssues":3,"license":null,"language":"JavaScript","lastCommitAt":"2026-09-21T13:51:51Z","createdAt":"2026-09-19T03:18:28Z","metadataFetchedAt":"2026-09-21T22:08:21.041Z","avatarUrl":"https://avatars.githubusercontent.com/u/59012?v=4","summarySource":"human-reviewed","claimStatus":"固定版本的公开源码已核对。本站未执行该项目，也未独立测量性能或安全性。","claimStatusEn":"Pinned public source was reviewed. This site did not execute the project or independently measure performance or safety.","plainSummaryJa":"各行が意味に合うかを Jev で採点する grep。AND/OR/NOT と言語をまたいだ検索ができる。","plainSummaryKo":"각 행이 어떤 의미에 맞는지 Jev로 채점하는 grep. AND/OR/NOT과 언어를 넘나드는 검색이 된다.","jevDecisionPointJa":"約 30 行をまとめ、行ごとに jev-latest へ Score または Noul を送り、意味に合うか聞く。","jevDecisionPointKo":"약 30행을 묶어 행마다 jev-latest에 Score 또는 Noul을 보내 해당 의미와 맞는지 묻는다.","highlightBenefitJa":"実行時依存のない意味 grep。クエリ文字列は TypeSafe に送られる。","highlightBenefitKo":"런타임 의존성이 없는 의미 grep. 질의 텍스트는 TypeSafe로 전송된다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T15:50:00.000Z","evidence":[{"url":"https://github.com/uehaj/jev-semgrep/blob/21120e9ee634defdb279b7243e03ea84aaeed2da/semgrep.mjs","note":"固定版本源码已核对；本站未执行该项目。"},{"url":"https://github.com/uehaj/jev-semgrep/blob/21120e9ee634defdb279b7243e03ea84aaeed2da/README.md","note":"固定版本源码已核对；本站未执行该项目。"},{"url":"https://github.com/uehaj/jev-semgrep/blob/21120e9ee634defdb279b7243e03ea84aaeed2da/LICENSE","note":"固定版本源码已核对；本站未执行该项目。"},{"url":"https://x.com/mohritaroh/status/2101325806119035332","note":"X discovery lead; integration is proven by the pinned GitHub files, not by the post."}],"pinned":false},{"id":"ddnim:jev-tweet-radar","name":"jev-tweet-radar","author":"DDnim","url":"https://github.com/DDnim/jev-tweet-radar","category":"Browser & OS Action","plainSummary":"Chrome 扩展对 X 时间线里每条帖子发一次 Jev Noul 请求，显示互动价值和可选标签概率。","plainSummaryEn":"A Chrome extension that scores each X timeline post with one Jev Noul batch for engagement value and optional tags.","jevDecisionPoint":"一次 System One 请求里对“是否值得互动”以及 spam、buzz、AI 等内容标签做 Noul。","jevDecisionPointEn":"One System One request asks Nouls for worth-engaging plus optional tags such as spam, buzz, and AI-ish.","highlightBenefit":"把时间线筛选做成可检查的概率，而不是生成评论文本。帖子正文会发到 TypeSafe。","highlightBenefitEn":"Turns timeline filtering into inspectable probabilities instead of generated commentary. Post text is sent to TypeSafe.","tags":["browser-automation","classification-ranking","typed-decisions"],"stars":1,"forks":0,"openIssues":0,"license":"MIT","language":"JavaScript","lastCommitAt":"2026-09-19T15:10:38Z","createdAt":"2026-09-19T13:24:22Z","metadataFetchedAt":"2026-09-21T22:08:24.102Z","avatarUrl":"https://avatars.githubusercontent.com/u/5253392?v=4","summarySource":"human-reviewed","claimStatus":"固定版本的公开源码已核对。本站未执行该项目，也未独立测量性能或安全性。","claimStatusEn":"Pinned public source was reviewed. This site did not execute the project or independently measure performance or safety.","plainSummaryJa":"X タイムラインの各投稿を 1 回の Jev Noul で採点し、交流価値と任意タグの確率を出す Chrome 拡張。","plainSummaryKo":"X 타임라인의 각 게시물을 Jev Noul 한 번으로 채점해 상호작용 가치와 선택 태그 확률을 보여주는 Chrome 확장.","jevDecisionPointJa":"1 回の System One で「交流する価値」と spam / buzz / AI などの Noul を聞く。","jevDecisionPointKo":"System One 요청 한 번에 ‘상호작용할 가치’와 spam, buzz, AI 같은 Noul을 묻는다.","highlightBenefitJa":"タイムライン選別を生成文ではなく確認できる確率にする。投稿本文は TypeSafe に送られる。","highlightBenefitKo":"타임라인 선별을 생성 댓글이 아니라 확인할 수 있는 확률로 만든다. 게시 본문은 TypeSafe로 전송된다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T15:50:00.000Z","evidence":[{"url":"https://github.com/DDnim/jev-tweet-radar/blob/b5057858ab5de6f7d665527a192f009eff8d0e76/background.js","note":"固定版本源码已核对；本站未执行该项目。"},{"url":"https://github.com/DDnim/jev-tweet-radar/blob/b5057858ab5de6f7d665527a192f009eff8d0e76/README.md","note":"固定版本源码已核对；本站未执行该项目。"},{"url":"https://x.com/WarlockTome/status/2101312284308172803","note":"X discovery lead; integration is proven by the pinned GitHub files, not by the post."}],"pinned":false},{"id":"thestral12:pr-sieve","name":"pr-sieve","author":"Thestral12","url":"https://github.com/Thestral12/pr-sieve","category":"CLI & Pipelines","plainSummary":"GitHub Action 把 `.jev.yml` 规则编译成 Jev 问题，按数值决定 fail、comment 或 pass。","plainSummaryEn":"A GitHub Action that compiles `.jev.yml` rules into Jev questions and fails, comments, or passes from the numbers.","jevDecisionPoint":"规则编成最多 12 个问题（`src/types.ts` 的 `MAX_JEV_QUESTIONS`）；`AKIA…` 与私钥装甲由 `src/redact.ts` 命中后，`src/pipeline.ts` 本地失败且不调用 Jev。","jevDecisionPointEn":"Rules become at most twelve questions (`MAX_JEV_QUESTIONS` in src/types.ts). AKIA and private-key armor from src/redact.ts fail locally in src/pipeline.ts with no Jev call.","highlightBenefit":"不写审查文、不给补丁、不自动批准；策略读自 base 分支配置。","highlightBenefitEn":"Does not write reviews, suggest patches, or approve. Policy is read from the base-ref config.","tags":["cli-git-gates","security-guardrails","typed-decisions"],"stars":1,"forks":0,"openIssues":0,"license":"MIT","language":"TypeScript","lastCommitAt":"2026-09-18T16:56:20Z","createdAt":"2026-09-18T16:41:32Z","metadataFetchedAt":"2026-09-21T22:08:27.087Z","avatarUrl":"https://avatars.githubusercontent.com/u/183735965?v=4","summarySource":"human-reviewed","claimStatus":"固定版本的公开源码已核对。本站未执行该项目，也未独立测量性能或安全性。","claimStatusEn":"Pinned public source was reviewed. This site did not execute the project or independently measure performance or safety.","plainSummaryJa":"`.jev.yml` の規則を Jev の質問にし、数値で fail / comment / pass を決める GitHub Action。","plainSummaryKo":"`.jev.yml` 규칙을 Jev 질문으로 바꿔 숫자로 fail, comment, pass를 정하는 GitHub Action.","jevDecisionPointJa":"規則は最大 12 問（src/types.ts の MAX_JEV_QUESTIONS）。AKIA と秘密鍵装甲は src/redact.ts が当て、src/pipeline.ts が Jev を呼ばず失敗する。","jevDecisionPointKo":"규칙은 최대 12개 질문(src/types.ts의 MAX_JEV_QUESTIONS). AKIA와 개인키 아머는 src/redact.ts가 잡고 src/pipeline.ts가 Jev를 부르지 않고 실패한다.","highlightBenefitJa":"レビュー文もパッチも自動承認もしない。方針は base 側の設定から読む。","highlightBenefitKo":"리뷰 문장, 패치, 자동 승인은 하지 않는다. 정책은 base 설정에서 읽는다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T15:50:00.000Z","evidence":[{"url":"https://github.com/Thestral12/pr-sieve/blob/9f37d439e37d99a145bc8367d982fe68df030e57/src/jev/client.ts","note":"固定版本源码已核对；本站未执行该项目。"},{"url":"https://github.com/Thestral12/pr-sieve/blob/9f37d439e37d99a145bc8367d982fe68df030e57/README.md","note":"固定版本源码已核对；本站未执行该项目。"},{"url":"https://x.com/scottyfost/status/2100999355486339348","note":"X discovery lead; integration is proven by the pinned GitHub files, not by the post."}],"pinned":false},{"id":"nitoba:questions","name":"questions","author":"nitoba","url":"https://github.com/nitoba/questions","category":"SDK & Decision Frameworks","plainSummary":"TypeScript 决策库：用 Zod 或原生问题描述判断，默认请求 TypeSafe Jev，也可改用 Vercel 或生成式适配器。","plainSummaryEn":"A TypeScript decision library that asks typed questions via Zod or native batches, defaulting to TypeSafe Jev, with optional Vercel or generative adapters.","jevDecisionPoint":"把 state 和 Choice / Score / Noul 发到 `/v1/systemone`，默认模型 `jev-latest`。","jevDecisionPointEn":"Posts state plus Choice, Score, or Noul questions to `/v1/systemone`, defaulting to jev-latest.","highlightBenefit":"同一套问题定义可以换 TypeSafe、Vercel 或生成式后端，不改调用代码。","highlightBenefitEn":"The same question definitions can target TypeSafe, Vercel, or a generative backend without rewriting callers.","tags":["multilanguage-sdk","typed-decisions"],"stars":5,"forks":1,"openIssues":0,"license":"MIT","language":"TypeScript","lastCommitAt":"2026-09-17T21:51:02Z","createdAt":"2026-09-17T09:34:38Z","metadataFetchedAt":"2026-09-21T22:08:30.166Z","avatarUrl":"https://avatars.githubusercontent.com/u/38570443?v=4","summarySource":"human-reviewed","claimStatus":"固定版本的公开源码已核对。本站未执行该项目，也未独立测量性能或安全性。","claimStatusEn":"Pinned public source was reviewed. This site did not execute the project or independently measure performance or safety.","plainSummaryJa":"Zod または素の質問で判断する TypeScript ライブラリ。既定は TypeSafe Jev で、Vercel や生成モデルにも差し替えられる。","plainSummaryKo":"Zod 또는 기본 질문으로 판단하는 TypeScript 라이브러리. 기본은 TypeSafe Jev이며 Vercel이나 생성 어댑터로 바꿀 수 있다.","jevDecisionPointJa":"state と Choice / Score / Noul を `/v1/systemone` へ送る。既定モデルは jev-latest。","jevDecisionPointKo":"state와 Choice, Score, Noul을 `/v1/systemone`으로 보낸다. 기본 모델은 jev-latest.","highlightBenefitJa":"同じ質問定義のまま TypeSafe、Vercel、生成バックエンドを切り替えられる。","highlightBenefitKo":"같은 질문 정의로 TypeSafe, Vercel, 생성 백엔드를 호출부 수정 없이 바꿀 수 있다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T15:50:00.000Z","evidence":[{"url":"https://github.com/nitoba/questions/blob/37501199437e3f3be11e185c157d30fa7a00601a/src/providers/typesafe.ts","note":"固定版本源码已核对；本站未执行该项目。"},{"url":"https://github.com/nitoba/questions/blob/37501199437e3f3be11e185c157d30fa7a00601a/src/providers/system-one.ts","note":"固定版本源码已核对；本站未执行该项目。"},{"url":"https://github.com/nitoba/questions/blob/37501199437e3f3be11e185c157d30fa7a00601a/README.md","note":"固定版本源码已核对；本站未执行该项目。"},{"url":"https://x.com/nito_b_a/status/2100949243485700547","note":"X discovery lead; integration is proven by the pinned GitHub files, not by the post."}],"pinned":false},{"id":"kentcdodds:kody","name":"kody","author":"kentcdodds","url":"https://github.com/kentcdodds/kody","category":"Data & Search","plainSummary":"可选的二段检索：先扩大混合召回，再用 Cloudflare Workers AI 上的 `typesafe/jev` Score 重排。","plainSummaryEn":"Optional second-stage search: widen the hybrid pool, then Score-rerank candidates with Workers AI typesafe/jev.","jevDecisionPoint":"对每个候选发 Score 问题，按分数重排并丢掉低分结果；模型 ID 为 `typesafe/jev`。","jevDecisionPointEn":"Sends a Score question per candidate, reorders and drops low scores; the model id is typesafe/jev.","highlightBenefit":"把 Jev 重排接到现有 MCP 搜索，而不是改写召回本身。许可证是 Fair Source FSL-1.1-ALv2，不是 OSI 开源。","highlightBenefitEn":"Attaches Jev rerank to existing MCP search instead of replacing retrieval. License is Fair Source FSL-1.1-ALv2, not OSI open source.","tags":["search-retrieval","mcp-integrations","classification-ranking"],"stars":662,"forks":65,"openIssues":26,"license":null,"language":"TypeScript","lastCommitAt":"2026-09-21T21:50:20Z","createdAt":"2026-03-18T23:30:18Z","metadataFetchedAt":"2026-09-21T22:08:33.035Z","avatarUrl":"https://avatars.githubusercontent.com/u/1500684?v=4","summarySource":"human-reviewed","claimStatus":"固定版本的公开源码已核对。许可证为 FSL-1.1-ALv2。本站未执行该项目，也未独立测量性能或安全性。","claimStatusEn":"Pinned public source was reviewed. The license is FSL-1.1-ALv2. This site did not execute the project or independently measure performance or safety.","plainSummaryJa":"任意の二段検索。ハイブリッド候補を広げたあと、Workers AI の typesafe/jev Score で並べ替える。","plainSummaryKo":"선택적 2단 검색: 혼합 후보를 넓힌 뒤 Workers AI typesafe/jev Score로 다시 정렬한다.","jevDecisionPointJa":"候補ごとに Score を送り、低分を落として並べ替える。モデル ID は typesafe/jev。","jevDecisionPointKo":"후보마다 Score를 보내 낮은 점수를 버리고 다시 정렬한다. 모델 ID는 typesafe/jev.","highlightBenefitJa":"既存の MCP 検索に Jev 再順位付けを足す。ライセンスは Fair Source FSL-1.1-ALv2 で OSI オープンソースではない。","highlightBenefitKo":"기존 MCP 검색에 Jev 재정렬을 붙인다. 라이선스는 Fair Source FSL-1.1-ALv2이며 OSI 오픈소스가 아니다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。ライセンスは FSL-1.1-ALv2。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 라이선스는 FSL-1.1-ALv2입니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"custom","sourceReviewedAt":"2026-09-19T15:50:00.000Z","evidence":[{"url":"https://github.com/kentcdodds/kody/blob/ed9275186dc817530dd2289e62c208ac3809391f/packages/worker/src/mcp/tools/search-jev-rerank.ts","note":"固定版本源码已核对；本站未执行该项目。"},{"url":"https://github.com/kentcdodds/kody/blob/ed9275186dc817530dd2289e62c208ac3809391f/LICENSE","note":"固定版本源码已核对；本站未执行该项目。"},{"url":"https://x.com/kodykoala/status/2100943346575253515","note":"X discovery lead; integration is proven by the pinned GitHub files, not by the post."}],"pinned":false,"topics":["agents","ai-assistant","cloudflare-workers","code-mode","mcp","personal-assistant"]},{"id":"jandalhuysen:jev-clash-royale-test","name":"jev-clash-royale-test","author":"JanDalhuysen","url":"https://github.com/JanDalhuysen/jev-clash-royale-test","category":"High-Frequency & Simulation","plainSummary":"Clash Royale 风格沙盒里，Jev 在一次请求中决定是否出牌、出哪张、哪条路和站位深度。","plainSummaryEn":"A Clash Royale-style sandbox whose Jev bot decides play-or-hold, card, lane, and depth in one System One call.","jevDecisionPoint":"`should_play` Noul 加上手牌 Choice、路线 Choice 和站位 Choice，一次 `systemOne`。","jevDecisionPointEn":"One systemOne call with a should_play Noul plus card, lane, and depth Choices.","highlightBenefit":"把出牌选择做成可检查的概率；package.json 写 ISC，仓库没有 LICENSE 文件。","highlightBenefitEn":"Makes play choices inspectable probabilities. package.json says ISC; there is no LICENSE file.","tags":["games-simulation","typed-decisions"],"stars":0,"forks":0,"openIssues":0,"license":null,"language":"JavaScript","lastCommitAt":"2026-09-19T11:10:48Z","createdAt":"2026-09-19T10:57:30Z","metadataFetchedAt":"2026-09-21T22:08:36.054Z","avatarUrl":"https://avatars.githubusercontent.com/u/58688790?v=4","summarySource":"human-reviewed","claimStatus":"固定版本的公开源码已核对。package.json 写明 ISC，但没有 LICENSE 文件。本站未执行该项目，也未独立测量性能或安全性。","claimStatusEn":"Pinned public source was reviewed. package.json lists ISC but there is no LICENSE file. This site did not execute the project or independently measure performance or safety.","plainSummaryJa":"Clash Royale 風サンドボックス。Jev が 1 回の要求で出す/待つ、カード、レーン、配置深さを決める。","plainSummaryKo":"Clash Royale 풍 샌드박스. Jev가 한 요청에서 낼지, 어떤 카드, 어느 레인, 배치 깊이를 정한다.","jevDecisionPointJa":"should_play の Noul とカード・レーン・深さの Choice を 1 回の systemOne で聞く。","jevDecisionPointKo":"should_play Noul과 카드, 레인, 깊이 Choice를 systemOne 한 번에 묻는다.","highlightBenefitJa":"出す選択を確認できる確率にする。package.json は ISC だが LICENSE ファイルはない。","highlightBenefitKo":"출수 선택을 확인할 수 있는 확률로 만든다. package.json은 ISC지만 LICENSE 파일은 없다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。package.json は ISC ですが LICENSE ファイルはありません。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. package.json은 ISC이지만 LICENSE 파일은 없습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"unconfirmed","sourceReviewedAt":"2026-09-19T16:30:00.000Z","evidence":[{"url":"https://github.com/JanDalhuysen/jev-clash-royale-test/blob/d25d0c7b4325a2c2f1449d40e1d834ff50069e21/jev_bot_loop.js","note":"固定版本源码已核对；本站未执行该项目。"},{"url":"https://github.com/JanDalhuysen/jev-clash-royale-test/blob/d25d0c7b4325a2c2f1449d40e1d834ff50069e21/README.md","note":"固定版本源码已核对；本站未执行该项目。"},{"url":"https://x.com/JanDalhuysen/status/2101334340353318993","note":"X discovery lead; integration is proven by the pinned GitHub files, not by the post."}],"pinned":false},{"id":"jaewgwon:jevis","name":"jevis","author":"jaewgwon","url":"https://github.com/jaewgwon/jevis","category":"Browser & OS Action","plainSummary":"Flutter integration_test 包：注册允许的 UI 动作，Jev 根据当前界面选下一步并判断目标是否达成。","plainSummaryEn":"A Flutter integration_test package that registers allowed UI actions and lets Jev pick the next action and whether the goal is done.","jevDecisionPoint":"默认 `jev-latest` 向 `/v1/systemone` POST：目标是 Noul，下一步是已注册动作上的 Choice。","jevDecisionPointEn":"POSTs to /v1/systemone as jev-latest: a goal Noul and a Choice over registered Flutter actions.","highlightBenefit":"把自然语言测试收成封闭动作表上的选择，而不是让模型自由点屏幕。","highlightBenefitEn":"Turns natural-language tests into choices over a closed action list instead of free-form tapping.","tags":["desktop-os","typed-decisions"],"stars":1,"forks":0,"openIssues":0,"license":"Apache-2.0","language":"Dart","lastCommitAt":"2026-09-19T09:00:06Z","createdAt":"2026-09-19T04:41:49Z","metadataFetchedAt":"2026-09-21T22:08:39.114Z","avatarUrl":"https://avatars.githubusercontent.com/u/40692132?v=4","summarySource":"human-reviewed","claimStatus":"固定版本的公开源码已核对。本站未执行该项目，也未独立测量性能或安全性。","claimStatusEn":"Pinned public source was reviewed. This site did not execute the project or independently measure performance or safety.","plainSummaryJa":"Flutter の integration_test 用。許可した UI 操作を登録し、Jev が次の操作と目標達成を選ぶ。","plainSummaryKo":"Flutter integration_test 패키지. 허용 UI 동작을 등록하고 Jev가 다음 동작과 목표 달성을 고른다.","jevDecisionPointJa":"既定 jev-latest で /v1/systemone に POST。目標は Noul、次の操作は登録済み Choice。","jevDecisionPointKo":"기본 jev-latest로 /v1/systemone에 POST. 목표는 Noul, 다음 동작은 등록된 Choice.","highlightBenefitJa":"自然言語テストを、自由なタップではなく閉じた操作表からの選択にする。","highlightBenefitKo":"자연어 테스트를 자유 탭이 아니라 닫힌 동작 표에서의 선택으로 만든다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T16:30:00.000Z","evidence":[{"url":"https://github.com/jaewgwon/jevis/blob/ab5630590288706ce56bb5665680358fe0bd95e8/lib/src/jevis_policy.dart","note":"固定版本源码已核对；本站未执行该项目。"},{"url":"https://github.com/jaewgwon/jevis/blob/ab5630590288706ce56bb5665680358fe0bd95e8/README.md","note":"固定版本源码已核对；本站未执行该项目。"},{"url":"https://github.com/jaewgwon/jevis/blob/ab5630590288706ce56bb5665680358fe0bd95e8/LICENSE","note":"固定版本源码已核对；本站未执行该项目。"},{"url":"https://x.com/jake_gwon/status/2101194639948787818","note":"X discovery lead; integration is proven by the pinned GitHub files, not by the post."}],"pinned":false,"topics":["e2e","flutter","mobile","test","testing"]},{"id":"hnegishi:typesafe-ai-ruby","name":"typesafe-ai-ruby","author":"hnegishi","url":"https://github.com/hnegishi/typesafe-ai-ruby","category":"SDK & Decision Frameworks","plainSummary":"无第三方运行时依赖的 Ruby 客户端，把 Choice / Score / Noul 发到 TypeSafe System One。","plainSummaryEn":"A stdlib-only Ruby client that sends Choice, Score, or Noul questions to TypeSafe System One.","jevDecisionPoint":"`system_one` POST 到 `/v1/systemone`，默认 `https://api.typesafe.ai` 与 `jev-latest`。","jevDecisionPointEn":"system_one POSTs to /v1/systemone, defaulting to https://api.typesafe.ai and jev-latest.","highlightBenefit":"给 Ruby 程序一条标准库路径去问 Jev，不必再包一层生成式 SDK。","highlightBenefitEn":"Gives Ruby programs a stdlib path to ask Jev without wrapping a generative SDK.","tags":["multilanguage-sdk","typed-decisions"],"stars":0,"forks":0,"openIssues":0,"license":"MIT","language":"Ruby","lastCommitAt":"2026-09-20T13:37:43Z","createdAt":"2026-09-19T02:44:15Z","metadataFetchedAt":"2026-09-21T22:08:42.066Z","avatarUrl":"https://avatars.githubusercontent.com/u/30515665?v=4","summarySource":"human-reviewed","claimStatus":"固定版本的公开源码已核对。本站未执行该项目，也未独立测量性能或安全性。","claimStatusEn":"Pinned public source was reviewed. This site did not execute the project or independently measure performance or safety.","plainSummaryJa":"第三者 runtime に依存しない Ruby クライアント。Choice / Score / Noul を TypeSafe System One へ送る。","plainSummaryKo":"서드파티 런타임 없이 Choice, Score, Noul을 TypeSafe System One으로 보내는 Ruby 클라이언트.","jevDecisionPointJa":"system_one が /v1/systemone へ POST。既定は https://api.typesafe.ai と jev-latest。","jevDecisionPointKo":"system_one이 /v1/systemone으로 POST. 기본은 https://api.typesafe.ai와 jev-latest.","highlightBenefitJa":"生成 SDK を挟まず、Ruby から Jev に聞く標準ライブラリ経路を提供する。","highlightBenefitKo":"생성 SDK를 감싸지 않고 Ruby에서 Jev를 묻는 표준 라이브러리 경로를 제공한다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T16:30:00.000Z","evidence":[{"url":"https://github.com/hnegishi/typesafe-ai-ruby/blob/4ab8c243fbd9e9d5526b59186595d0534785d963/lib/typesafe/client.rb","note":"固定版本源码已核对；本站未执行该项目。"},{"url":"https://github.com/hnegishi/typesafe-ai-ruby/blob/4ab8c243fbd9e9d5526b59186595d0534785d963/lib/typesafe/constants.rb","note":"固定版本源码已核对；本站未执行该项目。"},{"url":"https://github.com/hnegishi/typesafe-ai-ruby/blob/4ab8c243fbd9e9d5526b59186595d0534785d963/README.md","note":"固定版本源码已核对；本站未执行该项目。"},{"url":"https://github.com/hnegishi/typesafe-ai-ruby/blob/4ab8c243fbd9e9d5526b59186595d0534785d963/LICENSE.txt","note":"固定版本源码已核对；本站未执行该项目。"},{"url":"https://x.com/hik4_n/status/2101173099425345927","note":"X discovery lead; integration is proven by the pinned GitHub files, not by the post."}],"pinned":false},{"id":"kuldeepsinh19:jev-decision-gateway","name":"jev-decision-gateway","author":"kuldeepsinh19","url":"https://github.com/kuldeepsinh19/jev-decision-gateway","category":"Routing & Cost Optimization","plainSummary":"把是否继续、用哪个工具、要不要校验交给 Jev，只有策略允许时才调用生成式 LLM。","plainSummaryEn":"A gateway that asks Jev continue / tool / verify questions and invokes a generative LLM only when policy says generation is needed.","jevDecisionPoint":"`TypeSafeClient.systemOne` 回答策略问题；适配器根据答案决定是否把请求交给生成模型。","jevDecisionPointEn":"TypeSafeClient.systemOne answers policy questions; the adapter then decides whether a generative model runs.","highlightBenefit":"把贵的生成调用挡在 Jev 门禁之后；作者给出的省调用数字未经本站复测。","highlightBenefitEn":"Places expensive generation behind a Jev gate. Author-reported call savings were not retested here.","tags":["llm-routing-cost","typed-decisions","coding-agents"],"stars":0,"forks":0,"openIssues":0,"license":"MIT","language":"TypeScript","lastCommitAt":"2026-09-19T03:49:28Z","createdAt":"2026-09-19T03:51:46Z","metadataFetchedAt":"2026-09-21T22:08:45.065Z","avatarUrl":"https://avatars.githubusercontent.com/u/125124603?v=4","summarySource":"human-reviewed","claimStatus":"固定版本的公开源码已核对。本站未执行该项目，也未独立测量性能或安全性。","claimStatusEn":"Pinned public source was reviewed. This site did not execute the project or independently measure performance or safety.","plainSummaryJa":"続けるか、どのツールか、検証するかを Jev に聞き、方針が許すときだけ生成 LLM を呼ぶ。","plainSummaryKo":"계속할지, 어떤 도구인지, 검증할지를 Jev에 묻고 정책이 허용할 때만 생성 LLM을 부른다.","jevDecisionPointJa":"TypeSafeClient.systemOne が方針の質問に答え、アダプタが生成モデルを呼ぶかを決める。","jevDecisionPointKo":"TypeSafeClient.systemOne이 정책 질문에 답하고, 어댑터가 생성 모델을 부를지 정한다.","highlightBenefitJa":"高価な生成呼び出しを Jev の門の後ろに置く。作者の削減数は未再測。","highlightBenefitKo":"비싼 생성 호출을 Jev 문 뒤에 둔다. 저자가 적은 절감 숫자는 여기서 재측정하지 않았다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T16:30:00.000Z","evidence":[{"url":"https://github.com/kuldeepsinh19/jev-decision-gateway/blob/67111792e840e150239d12337a1d939be68be0e6/packages/core/src/jev/client.ts","note":"固定版本源码已核对；本站未执行该项目。"},{"url":"https://github.com/kuldeepsinh19/jev-decision-gateway/blob/67111792e840e150239d12337a1d939be68be0e6/packages/core/src/jev/adapter.ts","note":"固定版本源码已核对；本站未执行该项目。"},{"url":"https://github.com/kuldeepsinh19/jev-decision-gateway/blob/67111792e840e150239d12337a1d939be68be0e6/README.md","note":"固定版本源码已核对；本站未执行该项目。"},{"url":"https://github.com/kuldeepsinh19/jev-decision-gateway/blob/67111792e840e150239d12337a1d939be68be0e6/LICENSE","note":"固定版本源码已核对；本站未执行该项目。"},{"url":"https://x.com/Kuldeepzzz/status/2101166296973119555","note":"X discovery lead; integration is proven by the pinned GitHub files, not by the post."}],"pinned":false},{"id":"thruwire:foreman","name":"foreman","author":"thruwire","url":"https://github.com/thruwire/foreman","category":"CLI & Pipelines","plainSummary":"独立监督环读取编码工人的 diff、日志和测试，用 Jev Noul 判断卡住、跑偏、该验证，再由 Python 策略干预。","plainSummaryEn":"An independent supervisor loop that reads worker diffs, logs and tests, asks Jev Nouls about stuck / off-track / verify, then applies a Python policy.","jevDecisionPoint":"`AsyncTypeSafeClient.system_one`，默认 `jev-latest`，对监督问题发 Noul。","jevDecisionPointEn":"AsyncTypeSafeClient.system_one with default jev-latest sends Noul questions for supervision.","highlightBenefit":"监督不代替工人写代码；与目录里的 `Shifty-Eye-Games/foreman-jev` 不是同一个仓库。","highlightBenefitEn":"Supervision does not write the worker's code. Distinct from cataloged Shifty-Eye-Games/foreman-jev.","tags":["coding-agents","typed-decisions","cli-git-gates"],"stars":466,"forks":34,"openIssues":3,"license":"MIT","language":"Python","lastCommitAt":"2026-09-20T13:48:21Z","createdAt":"2026-09-17T11:46:13Z","metadataFetchedAt":"2026-09-21T22:08:48.042Z","avatarUrl":"https://avatars.githubusercontent.com/u/214398454?v=4","summarySource":"human-reviewed","claimStatus":"固定版本的公开源码已核对。本站未执行该项目，也未独立测量性能或安全性。","claimStatusEn":"Pinned public source was reviewed. This site did not execute the project or independently measure performance or safety.","plainSummaryJa":"作業者の diff・ログ・テストを読み、Jev の Noul で停滞・逸脱・検証を判断し、Python 方針で介入する監督ループ。","plainSummaryKo":"작업자의 diff, 로그, 테스트를 읽고 Jev Noul로 멈춤, 이탈, 검증을 판단한 뒤 Python 정책으로 개입하는 감독 루프.","jevDecisionPointJa":"既定 jev-latest の AsyncTypeSafeClient.system_one が監督用 Noul を送る。","jevDecisionPointKo":"기본 jev-latest의 AsyncTypeSafeClient.system_one이 감독용 Noul을 보낸다.","highlightBenefitJa":"監督は作業者のコードを書かない。掲載済み Shifty-Eye-Games/foreman-jev とは別リポジトリ。","highlightBenefitKo":"감독은 작업자 코드를 쓰지 않는다. 수록된 Shifty-Eye-Games/foreman-jev와는 다른 저장소다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T16:30:00.000Z","evidence":[{"url":"https://github.com/thruwire/foreman/blob/209182dac7a3467033fd093ab4ca47d21279984a/src/foreman/foreman/jev.py","note":"固定版本源码已核对；本站未执行该项目。"},{"url":"https://github.com/thruwire/foreman/blob/209182dac7a3467033fd093ab4ca47d21279984a/README.md","note":"固定版本源码已核对；本站未执行该项目。"},{"url":"https://github.com/thruwire/foreman/blob/209182dac7a3467033fd093ab4ca47d21279984a/LICENSE","note":"固定版本源码已核对；本站未执行该项目。"},{"url":"https://x.com/JoshARosen/status/2100573432089866717","note":"X discovery lead; integration is proven by the pinned GitHub files, not by the post."}],"pinned":false},{"id":"stas4000:jev-clerk","name":"jev-clerk","author":"stas4000","url":"https://github.com/stas4000/jev-clerk","category":"Browser & OS Action","plainSummary":"在 macOS 桌面上把供应商发票录入会计软件：Jev 每步从封闭动作表里选点击对象，深度模型只改剧本。","plainSummaryEn":"A macOS desktop clerk that books supplier invoices: Jev picks each click from a closed action list; a deep model only rewrites the playbook.","jevDecisionPoint":"POST `https://api.typesafe.ai/v1/systemone`，默认 `jev-latest`，每步做封闭动作 Choice。","jevDecisionPointEn":"POSTs https://api.typesafe.ai/v1/systemone as jev-latest with a closed action Choice each step.","highlightBenefit":"屏幕动作由 Jev 的封闭选择驱动；作者演示数字未经本站复测。GitHub SPDX 为空，LICENSE 文件是 MIT。","highlightBenefitEn":"Screen actions follow Jev's closed choices. Author demo numbers were not retested here. GitHub SPDX is empty; the LICENSE file is MIT.","tags":["desktop-os","domain-workflows","typed-decisions"],"stars":6,"forks":0,"openIssues":0,"license":null,"language":"Python","lastCommitAt":"2026-09-19T14:49:54Z","createdAt":"2026-09-19T14:22:55Z","metadataFetchedAt":"2026-09-21T22:08:51.131Z","avatarUrl":"https://avatars.githubusercontent.com/u/16101773?v=4","summarySource":"human-reviewed","claimStatus":"固定版本的公开源码已核对。本站未执行该项目，也未独立测量性能或安全性。","claimStatusEn":"Pinned public source was reviewed. This site did not execute the project or independently measure performance or safety.","plainSummaryJa":"macOS で仕入先請求を会計ソフトへ入力する。Jev が閉じた操作表からクリック先を選び、深いモデルは台本だけ直す。","plainSummaryKo":"macOS에서 공급 인보이스를 회계 앱에 입력한다. Jev가 닫힌 동작 표에서 클릭을 고르고, 깊은 모델은 대본만 고친다.","jevDecisionPointJa":"既定 jev-latest で /v1/systemone に POST し、各ステップで閉じた操作 Choice を聞く。","jevDecisionPointKo":"기본 jev-latest로 /v1/systemone에 POST하고 단계마다 닫힌 동작 Choice를 묻는다.","highlightBenefitJa":"画面操作は Jev の閉じた選択に従う。作者のデモ数値は未再測。GitHub SPDX は空で、LICENSE は MIT。","highlightBenefitKo":"화면 동작은 Jev의 닫힌 선택을 따른다. 저자 데모 숫자는 재측정하지 않았다. GitHub SPDX는 비어 있고 LICENSE는 MIT다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T16:30:00.000Z","evidence":[{"url":"https://github.com/stas4000/jev-clerk/blob/b471384e56bd9cec8f56ba8415f69b1006ce43e9/jev_clerk/jev.py","note":"固定版本源码已核对；本站未执行该项目。"},{"url":"https://github.com/stas4000/jev-clerk/blob/b471384e56bd9cec8f56ba8415f69b1006ce43e9/README.md","note":"固定版本源码已核对；本站未执行该项目。"},{"url":"https://github.com/stas4000/jev-clerk/blob/b471384e56bd9cec8f56ba8415f69b1006ce43e9/LICENSE","note":"固定版本源码已核对；本站未执行该项目。"},{"url":"https://x.com/stas_sorokin_/status/2101335212131012644","note":"X discovery lead; integration is proven by the pinned GitHub files, not by the post."}],"pinned":false},{"id":"florian-hoenicke:jev-gpt","name":"jev-gpt","author":"florian-hoenicke","url":"https://github.com/florian-hoenicke/jev-gpt","category":"High-Frequency & Simulation","plainSummary":"用级联 Choice 把 Jev 当成逐词分类器，在词树上逐层选出下一个词。","plainSummaryEn":"Cascaded Choice questions that make Jev pick the next word from a word tree instead of generating text.","jevDecisionPoint":"POST `/v1/systemone`，`jev-latest`，每层一个 `type: choice` 问题。","jevDecisionPointEn":"POSTs /v1/systemone as jev-latest with one type:choice question per tree level.","highlightBenefit":"展示 Jev 不生成长文时如何拼出短句；仓库没有 LICENSE 文件。","highlightBenefitEn":"Shows how Jev can assemble short text without generating prose. No LICENSE file is present.","tags":["typed-decisions","creative-multimedia"],"stars":2,"forks":1,"openIssues":0,"license":null,"language":"Python","lastCommitAt":"2026-09-19T15:17:02Z","createdAt":"2026-09-19T15:07:43Z","metadataFetchedAt":"2026-09-21T22:08:54.087Z","avatarUrl":"https://avatars.githubusercontent.com/u/11627845?v=4","summarySource":"human-reviewed","claimStatus":"固定版本的公开源码已核对。仓库没有 LICENSE 文件。本站未执行该项目，也未独立测量性能或安全性。","claimStatusEn":"Pinned public source was reviewed. There is no LICENSE file. This site did not execute the project or independently measure performance or safety.","plainSummaryJa":"段々の Choice で Jev を単語分類器にし、語彙木から次の語を選ぶ。","plainSummaryKo":"연쇄 Choice로 Jev를 단어 분류기로 써서 단어 나무에서 다음 단어를 고른다.","jevDecisionPointJa":"jev-latest で /v1/systemone に POST し、階層ごとに type:choice を 1 問出す。","jevDecisionPointKo":"jev-latest로 /v1/systemone에 POST하고 계층마다 type:choice 질문 하나를 낸다.","highlightBenefitJa":"長文を生成せず短い文を組み立てる例。LICENSE ファイルはない。","highlightBenefitKo":"긴 글을 생성하지 않고 짧은 문장을 조립하는 예. LICENSE 파일은 없다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。LICENSE ファイルはありません。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. LICENSE 파일이 없습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"unconfirmed","sourceReviewedAt":"2026-09-19T16:30:00.000Z","evidence":[{"url":"https://github.com/florian-hoenicke/jev-gpt/blob/25751a7f6a87ee31ee4ed2eb381fd746ca59c645/generate.py","note":"固定版本源码已核对；本站未执行该项目。"},{"url":"https://github.com/florian-hoenicke/jev-gpt/blob/25751a7f6a87ee31ee4ed2eb381fd746ca59c645/README.md","note":"固定版本源码已核对；本站未执行该项目。"},{"url":"https://x.com/florianhoenicke/status/2101333502369083690","note":"X discovery lead; integration is proven by the pinned GitHub files, not by the post."}],"pinned":false},{"id":"dotnetvibecoderz:vibe_sdk","name":"TypeSafeSDK","author":"DotNetVibeCoderz","url":"https://github.com/DotNetVibeCoderz/Vibe_SDK","category":"SDK & Decision Frameworks","plainSummary":"非官方 .NET 客户端向 TypeSafe `/v1/systemone` 发送 state 与 typed questions；父仓库还混有与 Jev 无关的 SDK。","plainSummaryEn":"An unofficial .NET client that POSTs state and typed questions to TypeSafe /v1/systemone. The parent repo also contains unrelated SDK dumps.","jevDecisionPoint":"`TypeSafeClient.SystemOneAsync` POST `{Endpoint}/v1/systemone`，默认 `https://api.typesafe.ai` 与 `jev-latest`。","jevDecisionPointEn":"TypeSafeClient.SystemOneAsync POSTs {Endpoint}/v1/systemone, defaulting to https://api.typesafe.ai and jev-latest.","highlightBenefit":"给 .NET 一条 HTTP 路径问 Jev；不要把父仓库里的其他 SDK 当成 Jev 集成。与已收录的 `saibimajdi/typesafeai-dotnet-sdk` 不是同一仓库。","highlightBenefitEn":"Gives .NET an HTTP path to ask Jev. Other SDKs in the parent dump are not Jev integrations. Distinct from cataloged saibimajdi/typesafeai-dotnet-sdk.","tags":["multilanguage-sdk","typed-decisions"],"stars":0,"forks":0,"openIssues":0,"license":"MIT","language":"C#","lastCommitAt":"2026-09-19T12:21:39Z","createdAt":"2026-08-01T13:24:46Z","metadataFetchedAt":"2026-09-21T22:08:57.179Z","avatarUrl":"https://avatars.githubusercontent.com/u/257091371?v=4","summarySource":"human-reviewed","claimStatus":"固定版本的公开源码已核对。本站未执行该项目，也未独立测量性能或安全性。","claimStatusEn":"Pinned public source was reviewed. This site did not execute the project or independently measure performance or safety.","plainSummaryJa":"非公式 .NET クライアントが TypeSafe /v1/systemone に state と typed questions を POST する。親リポジトリには Jev と無関係な SDK も混ざる。","plainSummaryKo":"비공식 .NET 클라이언트가 TypeSafe /v1/systemone에 state와 typed questions를 POST한다. 부모 저장소에는 Jev와 무관한 SDK도 섞여 있다.","jevDecisionPointJa":"TypeSafeClient.SystemOneAsync が {Endpoint}/v1/systemone へ POST。既定は https://api.typesafe.ai と jev-latest。","jevDecisionPointKo":"TypeSafeClient.SystemOneAsync가 {Endpoint}/v1/systemone으로 POST. 기본은 https://api.typesafe.ai와 jev-latest.","highlightBenefitJa":".NET から Jev に聞く HTTP 経路。親ダンプ内の他 SDK は Jev 連携ではない。掲載済み saibimajdi/typesafeai-dotnet-sdk とは別。","highlightBenefitKo":".NET에서 Jev를 묻는 HTTP 경로. 부모 덤프의 다른 SDK는 Jev 연동이 아니다. 수록된 saibimajdi/typesafeai-dotnet-sdk와는 다르다.","claimStatusJa":"固定バージョンの公開ソースを確認済みです。当サイトによる実行、性能測定、安全性監査は行っていません。","claimStatusKo":"고정 버전의 공개 소스를 확인했습니다. 이 사이트에서 직접 실행하거나 성능을 측정하거나 보안을 감사하지는 않았습니다.","licenseStatus":"declared","sourceReviewedAt":"2026-09-19T16:30:00.000Z","evidence":[{"url":"https://github.com/DotNetVibeCoderz/Vibe_SDK/blob/09c77029c00df6aa707b059e560512f17896b059/TypeSafeSDK/TypeSafeSdk/TypeSafeClient.cs","note":"固定版本源码已核对；本站未执行该项目。"},{"url":"https://github.com/DotNetVibeCoderz/Vibe_SDK/blob/09c77029c00df6aa707b059e560512f17896b059/TypeSafeSDK/README.md","note":"固定版本源码已核对；本站未执行该项目。"},{"url":"https://github.com/DotNetVibeCoderz/Vibe_SDK/blob/09c77029c00df6aa707b059e560512f17896b059/LICENSE","note":"固定版本源码已核对；本站未执行该项目。"},{"url":"https://x.com/Mifmasterz/status/2101296433408417841","note":"X discovery lead; integration is proven by the pinned GitHub files, not by the post."}],"pinned":false},{"id":"harshwasan:pi-jev-sentinel","name":"pi-jev-sentinel","author":"harshwasan","url":"https://github.com/harshwasan/pi-jev-sentinel","category":"Security & Guardrails","plainSummary":"这是一个 Pi coding-agent 扩展，用 Jev 检查工具调用的意图和风险，并筛查工具输出和回复中的注入指示，同时隐去密钥后再发送。","plainSummaryEn":"Open-source Pi coding-agent extension that uses Jev to check tool calls before they run, scan files for prompt injection, flag risky replies, and keep secrets out of what it sends.","jevDecisionPoint":"评估内容是否满足安全或证据要求，由本地策略放行或拦截。","jevDecisionPointEn":"Before each tool call, Jev receives the user's request (or pinned task), the proposed command, the agent's explanation, and recent conversation. It answers a Choice for intent (on task / off task / injected / needs more context) and a Score for risk (safe / needs approval / harmful). Local code turns these into run, ask the user, or warn. If Jev says more context would help, the extension adds it and asks again.","highlightBenefit":"给现有流程增加可检查的判断环节。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["security-guardrails","typed-decisions"],"stars":8,"forks":1,"openIssues":1,"license":"MIT","lastCommitAt":"2026-09-20T21:42:46Z","createdAt":"2026-09-19T16:40:22Z","metadataFetchedAt":"2026-09-21T22:09:03.035Z","avatarUrl":"https://avatars.githubusercontent.com/u/37144188?v=4","summarySource":"readme-extractive","claimStatus":"优先保留投稿者与仓库原文，缺失语言自动补充；自动检查仅确认 Jev 集成证据，未经本站运行或性能复测。","claimStatusEn":"Author and repository text is preserved where clear; missing languages are enriched automatically. Checks establish source-level Jev integration, not runtime, safety, or performance validation.","evidence":[{"url":"https://github.com/harshwasan/pi-jev-sentinel/blob/8a9ae145eaf165b9ff1c600ea0188a585aa12db5/README.md","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/harshwasan/pi-jev-sentinel/blob/8a9ae145eaf165b9ff1c600ea0188a585aa12db5/src/guard.ts","note":"固定版本的 Jev 集成与说明来源"}]},{"id":"shitianfang:jev-use","name":"jev-use","author":"shitianfang","url":"https://github.com/shitianfang/jev-use","category":"Context GC & Filter","plainSummary":"把 Claude Code / Codex / pi 中不需要输出文本的判断步骤交给 Jev 执行，需要写字或置信度不足的步骤按类型化契约退回 LLM。","plainSummaryEn":"Claude Code / Codex / pi plugin that hands agent steps needing no text output to Jev (TypeSafe's judgment model) — measured p50 ~230 ms and ~$0.02 per 1,000 judgments, with typed escalation back to the LLM","jevDecisionPoint":"`judge()`（第 35 行）把关于同一个 state 的多个 noul / choice / score 问题打成一次调用；`toVerdict()`（第 100~157 行）按原语分别处理 `noul`、`choice`、`score` 的答案与置信度。同文件的 `gate()`（第 170 行）是可选的 PreToolUse 门禁，只能 deny 或 ask。置信度不足或本就不该由 Jev 决定的步骤，通过类型化 escalation 契约（`writing`、`open_ended`、`oversized`、`unsure`、`unreachable`）退回 LLM，而不是让 Jev 猜。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"减少后续处理的冗余信息。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["context-compaction","classification-ranking","mcp-integrations"],"stars":14,"forks":1,"openIssues":1,"license":"MIT","lastCommitAt":"2026-09-20T16:52:50Z","createdAt":"2026-09-19T10:13:56Z","metadataFetchedAt":"2026-09-21T22:09:06.033Z","avatarUrl":"https://avatars.githubusercontent.com/u/61344989?v=4","summarySource":"readme-extractive","claimStatus":"优先保留投稿者与仓库原文，缺失语言自动补充；自动检查仅确认 Jev 集成证据，未经本站运行或性能复测。","claimStatusEn":"Author and repository text is preserved where clear; missing languages are enriched automatically. Checks establish source-level Jev integration, not runtime, safety, or performance validation.","evidence":[{"url":"https://github.com/shitianfang/jev-use/blob/56f3099ba5a9049195f931486450610ba1567514/README.md","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/shitianfang/jev-use/blob/56f3099ba5a9049195f931486450610ba1567514/README.zh-CN.md","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/shitianfang/jev-use/blob/56f3099ba5a9049195f931486450610ba1567514/src/backends/typesafe.ts","note":"固定版本的 Jev 集成与说明来源"}]},{"id":"replynodes:jev-web-analyzer","name":"jev-web-analyzer","author":"replynodes","url":"https://github.com/replynodes/jev-web-analyzer","category":"SDK & Decision Frameworks","plainSummary":"jev-web-analyzer：由业务代码定义问题，客户端负责提交 Jev 请求并解析结构化结果。","plainSummaryEn":"See what Jev thinks about your SaaS website — powered by ReplyNodes web context and Vercel AI Gateway.","jevDecisionPoint":"由业务代码定义问题，客户端负责提交 Jev 请求并解析结构化结果。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"在现有程序中复用接入代码，减少重复处理接口细节。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["multilanguage-sdk","typed-decisions"],"stars":3,"forks":0,"openIssues":0,"license":"Apache-2.0","lastCommitAt":"2026-09-21T14:38:11Z","createdAt":"2026-09-19T11:47:09Z","metadataFetchedAt":"2026-09-21T22:09:09.058Z","avatarUrl":"https://avatars.githubusercontent.com/u/307085583?v=4","summarySource":"readme-extractive","claimStatus":"优先保留投稿者与仓库原文，缺失语言自动补充；自动检查仅确认 Jev 集成证据，未经本站运行或性能复测。","claimStatusEn":"Author and repository text is preserved where clear; missing languages are enriched automatically. Checks establish source-level Jev integration, not runtime, safety, or performance validation.","evidence":[{"url":"https://github.com/replynodes/jev-web-analyzer/blob/bb50b94a540af98c4dae63fc6265f4513121353e/README.md","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/replynodes/jev-web-analyzer/blob/bb50b94a540af98c4dae63fc6265f4513121353e/app/api/analyze/route.test.ts","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/replynodes/jev-web-analyzer/blob/bb50b94a540af98c4dae63fc6265f4513121353e/app/api/analyze/route.ts","note":"固定版本的 Jev 集成与说明来源"}]},{"id":"harshpuri84:slopcheck-jev","name":"slopcheck-jev","author":"harshpuri84","url":"https://github.com/harshpuri84/slopcheck-jev","category":"CLI & Pipelines","plainSummary":"slopcheck-jev 在终端和自动化脚本中加入文本判断。","plainSummaryEn":"A prose linter that catches AI writing tells. Regex settles the 18 a pattern can settle. Jev takes the 15 that need reading, as 15 Nouls in one call, 604 ms median. It ships as a Claude Code `Stop` hook that scores Claude's own output after every turn and warns rather than blocks.","jevDecisionPoint":"对输入文本做分类或打分，交给本地规则继续处理。","jevDecisionPointEn":"**Deleting a feature made it more accurate.** The first build made a second Jev call to pin each tell to a sentence. Removing it took precision from 0.80 to 0.95 and latency from 1,365 ms to 575 ms. Jev returns typed answers rather than text, so a quote has to come from a second round of typed questions about lines, and that round was where every defect lived. The README documents the two intermediate designs that were worse.","highlightBenefit":"把语义判断接到已有的命令行工作流。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["cli-git-gates","typed-decisions"],"stars":0,"forks":0,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-19T21:15:09Z","createdAt":"2026-09-19T15:42:25Z","metadataFetchedAt":"2026-09-21T22:09:12.119Z","avatarUrl":"https://avatars.githubusercontent.com/u/234730463?v=4","summarySource":"readme-extractive","claimStatus":"优先保留投稿者与仓库原文，缺失语言自动补充；自动检查仅确认 Jev 集成证据，未经本站运行或性能复测。","claimStatusEn":"Author and repository text is preserved where clear; missing languages are enriched automatically. Checks establish source-level Jev integration, not runtime, safety, or performance validation.","evidence":[{"url":"https://github.com/harshpuri84/slopcheck-jev/blob/f37c440b84aaa09e0c2de5f441f19a32f8711781/README.md","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/harshpuri84/slopcheck-jev/blob/f37c440b84aaa09e0c2de5f441f19a32f8711781/slopcheck/gate.py","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/harshpuri84/slopcheck-jev/blob/f37c440b84aaa09e0c2de5f441f19a32f8711781/slopcheck/judge.py","note":"固定版本的 Jev 集成与说明来源"}]},{"id":"solido:jev_dart","name":"jev_dart","author":"Solido","url":"https://github.com/Solido/jev_dart","category":"SDK & Decision Frameworks","plainSummary":"这是 Jev 的纯 Dart 客户端，用于发送状态和类型化问题并返回可供代码分支的结构化答案，可用于 CLI、服务端和 Flutter 应用。","plainSummaryEn":"Jev Dart SDK to build cli, server and Flutter apps.","jevDecisionPoint":"Jev 根据输入文本在 billing/technical/other 之间做 Choice 分类，并用 Noul 判断是否紧急。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"在现有程序中复用接入代码，减少重复处理接口细节。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["multilanguage-sdk","typed-decisions"],"stars":1,"forks":0,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-21T00:40:45Z","createdAt":"2026-09-19T20:10:31Z","metadataFetchedAt":"2026-09-21T22:09:15.129Z","avatarUrl":"https://avatars.githubusercontent.com/u/1295961?v=4","summarySource":"readme-extractive","claimStatus":"优先保留投稿者与仓库原文，缺失语言自动补充；自动检查仅确认 Jev 集成证据，未经本站运行或性能复测。","claimStatusEn":"Author and repository text is preserved where clear; missing languages are enriched automatically. Checks establish source-level Jev integration, not runtime, safety, or performance validation.","evidence":[{"url":"https://github.com/Solido/jev_dart/blob/03e4d03fbfe2046c1b730364aac88d7df54e8a0f/README.md","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/Solido/jev_dart/blob/03e4d03fbfe2046c1b730364aac88d7df54e8a0f/lib/jev_dart.dart","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/Solido/jev_dart/blob/03e4d03fbfe2046c1b730364aac88d7df54e8a0f/bin/jev_dart.dart","note":"固定版本的 Jev 集成与说明来源"}]},{"id":"joslynsmall:fast-jev-compaction-pi","name":"fast-jev-compaction-pi","author":"joslynSmall","url":"https://github.com/joslynSmall/fast-jev-compaction-pi","category":"Context GC & Filter","plainSummary":"为 Pi 的上下文压缩筛选并原样保留关键工具证据，删除或截短过时的工具输出。","plainSummaryEn":"This Pi extension asks Jev whether each completed tool call and its full result should be kept, then locally retains, truncates, or drops verbatim tool evidence in the compaction summary.","jevDecisionPoint":"Jev 接收待压缩会话的任务目标、用户和助手文本、已完成工具调用的名称与参数、结果长度和错误标记，不接收工具输出正文。\n它对每个工具调用分别给出两个概率判断：是否保留“该调用及参数”，以及是否保留“完整原始结果”。本地代码再按阈值执行完整保留、截短结果或删除调用和结果；Jev 请求或解析失败时回退 Pi 原生摘要。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"把语义判断接到已有的命令行工作流。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["context-compaction","classification-ranking"],"stars":0,"forks":0,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-20T01:57:27Z","createdAt":"2026-09-20T01:25:20Z","metadataFetchedAt":"2026-09-21T22:09:18.105Z","avatarUrl":"https://avatars.githubusercontent.com/u/18278604?v=4","summarySource":"readme-extractive","claimStatus":"优先保留投稿者与仓库原文，缺失语言自动补充；自动检查仅确认 Jev 集成证据，未经本站运行或性能复测。","claimStatusEn":"Author and repository text is preserved where clear; missing languages are enriched automatically. Checks establish source-level Jev integration, not runtime, safety, or performance validation.","evidence":[{"url":"https://github.com/joslynSmall/fast-jev-compaction-pi/blob/e0086cbb70c684acf35cfe24261e65a708c9f045/README.md","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/joslynSmall/fast-jev-compaction-pi/blob/e0086cbb70c684acf35cfe24261e65a708c9f045/src/jev.ts","note":"固定版本的 Jev 集成与说明来源"}]},{"id":"socai-io:jev-social","name":"jev-social","author":"socai-io","url":"https://github.com/socai-io/jev-social","category":"Domain & Vertical Tools","plainSummary":"该项目由 Jev 逐步选择 Instagram、TikTok 和 LinkedIn 的搜索、打开主页或帖子、读取评论及下载视频等操作，并由 socai 在真实浏览器中执行后整理成带引用的报告。","plainSummaryEn":"Jev-powered Instagram, TikTok, and LinkedIn research: typed routing, real browser evidence, streamed post cards, video capture, and cited socai reports.","jevDecisionPoint":"Jev 在每一步从具体只读操作候选中选择平台路由和下一个要执行的搜索/打开/读取/下载或结束操作。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"把语义判断接到已有的命令行工作流。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["domain-workflows","typed-decisions"],"stars":40,"forks":5,"openIssues":6,"license":"MIT","lastCommitAt":"2026-09-21T17:05:53Z","createdAt":"2026-09-18T18:48:15Z","metadataFetchedAt":"2026-09-21T22:09:21.027Z","avatarUrl":"https://avatars.githubusercontent.com/u/291013799?v=4","summarySource":"readme-extractive","claimStatus":"优先保留投稿者与仓库原文，缺失语言自动补充；自动检查仅确认 Jev 集成证据，未经本站运行或性能复测。","claimStatusEn":"Author and repository text is preserved where clear; missing languages are enriched automatically. Checks establish source-level Jev integration, not runtime, safety, or performance validation.","evidence":[{"url":"https://github.com/socai-io/jev-social/blob/cde1ff8fe5793421fcedf7a78d34015fb28ca85f/README.md","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/socai-io/jev-social/blob/cde1ff8fe5793421fcedf7a78d34015fb28ca85f/bin/jev-social.js","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/socai-io/jev-social/blob/cde1ff8fe5793421fcedf7a78d34015fb28ca85f/public/app.js","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/socai-io/jev-social/blob/cde1ff8fe5793421fcedf7a78d34015fb28ca85f/src/actions.js","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/socai-io/jev-social/blob/cde1ff8fe5793421fcedf7a78d34015fb28ca85f/src/app.js","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/socai-io/jev-social/blob/cde1ff8fe5793421fcedf7a78d34015fb28ca85f/src/classifier.js","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/socai-io/jev-social/blob/cde1ff8fe5793421fcedf7a78d34015fb28ca85f/src/config.js","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/socai-io/jev-social/blob/cde1ff8fe5793421fcedf7a78d34015fb28ca85f/src/env.js","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/socai-io/jev-social/blob/cde1ff8fe5793421fcedf7a78d34015fb28ca85f/src/errors.js","note":"固定版本的 Jev 集成与说明来源"}]},{"id":"dimweaker:jev-libero","name":"jev-libero","author":"Dimweaker","url":"https://github.com/Dimweaker/jev-libero","category":"High-Frequency & Simulation","plainSummary":"在 LIBERO 仿真环境中，使用 Jev 分层选择机器人原子动作，结合可逆物理前视完成操作任务。支持 JSON 任务配置，并提供关闭微波炉、关闭抽屉的演示与完整运行记录。","plainSummaryEn":"Two LIBERO tasks, one control engine. Each demo loads its own JSON task definition. Videos follow simulation time, with decision and physics-preview waiting omitted.","jevDecisionPoint":"Jev 通过串行 Choice 调用选择意图 → 接触/运动方式 → 一个原子输入。引擎先用物理前视评价候选，随后执行 Jev 选中的输入，并根据新状态继续决策。动作集包括平移、旋转、开爪、合爪和保持，共 27 个输入。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"在连续交互中观察决策效果；频率依实际运行而定。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["games-simulation","typed-decisions","domain-workflows"],"stars":47,"forks":1,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-21T07:46:28Z","createdAt":"2026-09-19T18:45:40Z","metadataFetchedAt":"2026-09-21T22:09:24.079Z","avatarUrl":"https://avatars.githubusercontent.com/u/42308062?v=4","summarySource":"readme-extractive","claimStatus":"优先保留投稿者与仓库原文，缺失语言自动补充；自动检查仅确认 Jev 集成证据，未经本站运行或性能复测。","claimStatusEn":"Author and repository text is preserved where clear; missing languages are enriched automatically. Checks establish source-level Jev integration, not runtime, safety, or performance validation.","evidence":[{"url":"https://github.com/Dimweaker/jev-libero/blob/667b3bbeb0e40594e6a9993356ce84e70afc936d/README.md","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/Dimweaker/jev-libero/blob/667b3bbeb0e40594e6a9993356ce84e70afc936d/README.zh-CN.md","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/Dimweaker/jev-libero/blob/667b3bbeb0e40594e6a9993356ce84e70afc936d/src/jev_libero/client.py","note":"固定版本的 Jev 集成与说明来源"}]},{"id":"zeredy879:minojev","name":"minojev","author":"zeredy879","url":"https://github.com/zeredy879/minojev","category":"SDK & Decision Frameworks","plainSummary":"独立实现的 System-1 决策模型：一次前向传播输出 Choice、Noul(Boolean)、Score 的校准概率分布，全程零输出 token，可在笔记本 CPU 上离线训练。","plainSummaryEn":"Decisions, not tokens: minojev reads calibrated, typed probability distributions straight from hidden states in one forward pass — zero output tokens, fully reproducible on a laptop CPU.","jevDecisionPoint":"minojev 是独立开源复现（不调用 TypeSafe API）：自研 backbone + 决策头，从隐藏状态直接读取 Choice / Boolean(等价 Noul) / Score 分布，温度校准在 dev 集拟合。迷宫 Demo 中每步一次前向回答 9 个问题（1 个四向 Choice + 4 个安全 Noul + 1 个距离 Score），代码用安全概率做掩码后执行动作。附带六个业务领域（客服、审核、代码评审、退款、分诊、线索）的合成数据集与后训练流程。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"把语义判断接到已有的命令行工作流。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["multilanguage-sdk","typed-decisions","domain-workflows"],"stars":25,"forks":1,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-21T16:43:03Z","createdAt":"2026-09-18T06:25:36Z","metadataFetchedAt":"2026-09-21T22:09:27.040Z","avatarUrl":"https://avatars.githubusercontent.com/u/66517906?v=4","summarySource":"readme-extractive","claimStatus":"优先保留投稿者与仓库原文，缺失语言自动补充；自动检查仅确认 Jev 集成证据，未经本站运行或性能复测。","claimStatusEn":"Author and repository text is preserved where clear; missing languages are enriched automatically. Checks establish source-level Jev integration, not runtime, safety, or performance validation.","evidence":[{"url":"https://github.com/zeredy879/minojev/blob/36733a15440c80467386447fd41464ea46dd972d/README.md","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/zeredy879/minojev/blob/36733a15440c80467386447fd41464ea46dd972d/README.zh-CN.md","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/zeredy879/minojev/blob/36733a15440c80467386447fd41464ea46dd972d/src/minojev/model.py","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/zeredy879/minojev/blob/36733a15440c80467386447fd41464ea46dd972d/src/minojev/__init__.py","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/zeredy879/minojev/blob/36733a15440c80467386447fd41464ea46dd972d/src/minojev/backbone.py","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/zeredy879/minojev/blob/36733a15440c80467386447fd41464ea46dd972d/src/minojev/bench.py","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/zeredy879/minojev/blob/36733a15440c80467386447fd41464ea46dd972d/src/minojev/calibrate.py","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/zeredy879/minojev/blob/36733a15440c80467386447fd41464ea46dd972d/src/minojev/cli.py","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/zeredy879/minojev/blob/36733a15440c80467386447fd41464ea46dd972d/src/minojev/data.py","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/zeredy879/minojev/blob/36733a15440c80467386447fd41464ea46dd972d/src/minojev/encoding.py","note":"固定版本的 Jev 集成与说明来源"}]},{"id":"shantanugoel:ask-jev-skill","name":"ask-jev-skill","author":"shantanugoel","url":"https://github.com/shantanugoel/ask-jev-skill","category":"SDK & Decision Frameworks","plainSummary":"Hermes Agent 与通用智能体的 Jev 技能扩展：为代理提供封闭选项评估与置信度不足时的升级决策机制。","plainSummaryEn":"A skill for Hermes and other agents to query TypeSafe Jev for bounded option judgments and confidence escalations.","jevDecisionPoint":"为 Agent 工作流在多动作选择时调用 Jev 评估候选意图；低置信度时交由人工或上级模型处理。","jevDecisionPointEn":"Evaluates candidate intents for agent workflows; escalates uncertain choices to human review or higher-tier models.","highlightBenefit":"为自主智能体增加确定性的窄域决策保护，减少无限轮循环幻觉。","highlightBenefitEn":"Equips autonomous agents with deterministic narrow judgments, mitigating infinite-loop hallucinations.","tags":["multilanguage-sdk","typed-decisions"],"stars":37,"forks":1,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-17T11:15:56Z","createdAt":"2026-09-17T10:57:10Z","metadataFetchedAt":"2026-09-21T22:09:30.052Z","avatarUrl":"https://avatars.githubusercontent.com/u/390304?v=4","summarySource":"curated","claimStatus":"经由公共仓库代码审查确认真实 Jev 可执行集成证据；决策机制提取自实际源码，未经本站性能跑分测试。","claimStatusEn":"Verified with code-level Jev executable integration evidence; decision mechanisms extracted from source.","evidence":[{"url":"https://github.com/shantanugoel/ask-jev-skill/blob/72313d3c88f8dbff40a11ffe69656ff9843ea78e/README.md","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/shantanugoel/ask-jev-skill/blob/72313d3c88f8dbff40a11ffe69656ff9843ea78e/scripts/askjev.py","note":"固定版本的 Jev 集成与说明来源"}]},{"id":"sufianetaouil:every","name":"every","author":"sufianetaouil","url":"https://github.com/sufianetaouil/every","category":"Data & Search","plainSummary":"代码库全量函数语义问答工具：对代码库内每个函数发起是非问题提问，按 Noul 概率在秒级内排列出最相关的函数。","plainSummaryEn":"Ask a yes/no question of every function in a codebase. Ranked answers in seconds, for cents. Grep whose pattern is a question, powered by TypeSafe Jev.","jevDecisionPoint":"将自然语言问题转化为布尔/概率判断，遍历代码库函数计算匹配置信度，由本地策略排序呈现。","jevDecisionPointEn":"Converts natural language questions into boolean/probability queries across functions, ranking matches locally by confidence.","highlightBenefit":"以极低单次调用成本实现类似 grep 的自然语言函数级代码发现。","highlightBenefitEn":"Enables natural language function discovery across codebases at low cost, similar to an intelligent grep.","tags":["search-retrieval","classification-ranking"],"stars":5,"forks":0,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-17T06:02:19Z","createdAt":"2026-09-17T05:38:34Z","metadataFetchedAt":"2026-09-21T22:09:33.035Z","avatarUrl":"https://avatars.githubusercontent.com/u/15270674?v=4","summarySource":"curated","claimStatus":"经由公共仓库代码审查确认真实 Jev 可执行集成证据；决策机制提取自实际源码，未经本站性能跑分测试。","claimStatusEn":"Verified with code-level Jev executable integration evidence; decision mechanisms extracted from source.","evidence":[{"url":"https://github.com/sufianetaouil/every/blob/aaa72d582a831420dfd23a788e3bc948c798c248/README.md","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/sufianetaouil/every/blob/aaa72d582a831420dfd23a788e3bc948c798c248/reference/probe_jev.py","note":"固定版本的 Jev 集成与说明来源"}]},{"id":"phyous:tsai-civ2","name":"tsai-civ2","author":"phyous","url":"https://github.com/phyous/tsai-civ2","category":"High-Frequency & Simulation","plainSummary":"Jev 玩原版《文明 II》实验框架：在浏览器中运行经典游戏引擎，实时输出各行动的概率分布并做出决策。","plainSummaryEn":"An experimental harness where TypeSafe Jev plays classic Civilization II in a browser, computing live action probability distributions.","jevDecisionPoint":"每回合读取游戏地图与单位状态，由 Jev 评估城市建造、科技研发、外交姿态与行动走位。","jevDecisionPointEn":"Evaluates city builds, tech research, diplomatic postures, and unit moves per turn from game state.","highlightBenefit":"探索 Jev 实时概率分布在复杂策略博弈与长期规划环境下的落地表现。","highlightBenefitEn":"Demonstrates Jev real-time probability distributions applied to complex strategic planning and game loops.","tags":["games-simulation","typed-decisions"],"stars":1,"forks":1,"openIssues":0,"license":null,"lastCommitAt":"2026-09-18T03:20:46Z","createdAt":"2026-09-16T17:42:36Z","metadataFetchedAt":"2026-09-21T22:09:36.080Z","avatarUrl":"https://avatars.githubusercontent.com/u/716386?v=4","summarySource":"curated","claimStatus":"经由公共仓库代码审查确认真实 Jev 可执行集成证据；决策机制提取自实际源码，未经本站性能跑分测试。","claimStatusEn":"Verified with code-level Jev executable integration evidence; decision mechanisms extracted from source.","evidence":[{"url":"https://github.com/phyous/tsai-civ2/blob/dab97616d1b5ef2ae80d277076d14b8886eac009/README.md","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/phyous/tsai-civ2/blob/dab97616d1b5ef2ae80d277076d14b8886eac009/civ2/typesafe.py","note":"固定版本的 Jev 集成与说明来源"}]},{"id":"komikat:jev-bfs","name":"jev-bfs","author":"komikat","url":"https://github.com/komikat/jev-bfs","category":"Data & Search","plainSummary":"基于 Jev 引导的维基百科链接竞速寻路器：利用 Jev 评估出站链接相关性，在终端实时寻径两词条之间的最短路径。","plainSummaryEn":"Wikipedia link race pathfinder guided by Jev: assesses outbound links to navigate between two articles in real time.","jevDecisionPoint":"在 BFS 搜索树的每个分支点，由 Jev 对当前页面的出站链接进行语义启发式打分并剪枝。","jevDecisionPointEn":"Applies Jev semantic heuristic scoring and pruning to outbound links at each BFS exploration step.","highlightBenefit":"结合传统图搜索算法与概率决策，显著缩减维基百科多跳路径的搜索空间。","highlightBenefitEn":"Combines classic graph traversal with probabilistic judgments to prune the Wikipedia multi-hop search space.","tags":["search-retrieval","classification-ranking"],"stars":0,"forks":0,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-17T06:00:27Z","createdAt":"2026-09-17T06:00:48Z","metadataFetchedAt":"2026-09-21T22:09:39.072Z","avatarUrl":"https://avatars.githubusercontent.com/u/41486636?v=4","summarySource":"curated","claimStatus":"经由公共仓库代码审查确认真实 Jev 可执行集成证据；决策机制提取自实际源码，未经本站性能跑分测试。","claimStatusEn":"Verified with code-level Jev executable integration evidence; decision mechanisms extracted from source.","evidence":[{"url":"https://github.com/komikat/jev-bfs/blob/d83ef1199fabcb2922030a27e0db3e27f6cd759c/README.md","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/komikat/jev-bfs/blob/d83ef1199fabcb2922030a27e0db3e27f6cd759c/jev_bfs/__init__.py","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/komikat/jev-bfs/blob/d83ef1199fabcb2922030a27e0db3e27f6cd759c/jev_bfs/__main__.py","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/komikat/jev-bfs/blob/d83ef1199fabcb2922030a27e0db3e27f6cd759c/jev_bfs/cli.py","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/komikat/jev-bfs/blob/d83ef1199fabcb2922030a27e0db3e27f6cd759c/jev_bfs/core.py","note":"固定版本的 Jev 集成与说明来源"}]},{"id":"prismhq:jev-router","name":"jev-router","author":"prismhq","url":"https://github.com/prismhq/jev-router","category":"Routing & Cost Optimization","plainSummary":"基于 LiteLLM 与 Jev 构建的开源模型路由器：根据输入任务复杂度与上下文自动选择性价比最高的大模型。","plainSummaryEn":"An open-source LLM router built on LiteLLM and Jev: dynamically routes requests based on task complexity and context.","jevDecisionPoint":"通过 System-1 单次前向推理对输入 Prompt 评定推理需求档位，再由本地 LiteLLM 路由分发请求。","jevDecisionPointEn":"Evaluates reasoning requirements of input prompts via single forward pass, delegating dispatch to LiteLLM.","highlightBenefit":"无需复杂的启发式正则即可实现高吞吐的任务分级与模型降本调度。","highlightBenefitEn":"Achieves high-throughput prompt triage and cost optimization without brittle heuristic regexes.","tags":["llm-routing-cost","typed-decisions"],"stars":7,"forks":0,"openIssues":1,"license":"MIT","lastCommitAt":"2026-09-17T01:51:29Z","createdAt":"2026-09-17T01:51:32Z","metadataFetchedAt":"2026-09-21T22:09:42.086Z","avatarUrl":"https://avatars.githubusercontent.com/u/186116915?v=4","summarySource":"curated","claimStatus":"经由公共仓库代码审查确认真实 Jev 可执行集成证据；决策机制提取自实际源码，未经本站性能跑分测试。","claimStatusEn":"Verified with code-level Jev executable integration evidence; decision mechanisms extracted from source.","evidence":[{"url":"https://github.com/prismhq/jev-router/blob/583f0a1d1e0534cda3b6bbfa4b19aa1ec25d73a7/README.md","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/prismhq/jev-router/blob/583f0a1d1e0534cda3b6bbfa4b19aa1ec25d73a7/jev_router/__init__.py","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/prismhq/jev-router/blob/583f0a1d1e0534cda3b6bbfa4b19aa1ec25d73a7/jev_router/config.py","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/prismhq/jev-router/blob/583f0a1d1e0534cda3b6bbfa4b19aa1ec25d73a7/jev_router/deciders.py","note":"固定版本的 Jev 集成与说明来源"}]},{"id":"teyhouse:jev-secret-detection","name":"jev-secret-detection","author":"teyhouse","url":"https://github.com/teyhouse/jev-secret-detection","category":"Security & Guardrails","plainSummary":"利用 Jev 模型检验代码片段中的真实凭据泄露：评估小模型在代码安全门禁与敏感密钥识别中的表现。","plainSummaryEn":"Benchmark and tool evaluating how well TypeSafe Jev identifies real secret credentials in file snippets.","jevDecisionPoint":"将文件代码片段输入 Jev 判定是否存在硬编码密钥风险，返回布尔与置信评分。","jevDecisionPointEn":"Submits code snippets to Jev to evaluate hardcoded credential risk, returning boolean and confidence scores.","highlightBenefit":"验证低延迟无生成模型在 CI/CD pre-commit 安全静态扫描中的可行性与准确率。","highlightBenefitEn":"Evaluates the feasibility and precision of low-latency non-generative models in pre-commit security gates.","tags":["security-guardrails","typed-decisions"],"stars":0,"forks":0,"openIssues":0,"license":null,"lastCommitAt":"2026-09-18T06:50:34Z","createdAt":"2026-09-17T11:15:02Z","metadataFetchedAt":"2026-09-21T22:09:45.063Z","avatarUrl":"https://avatars.githubusercontent.com/u/828439?v=4","summarySource":"curated","claimStatus":"经由公共仓库代码审查确认真实 Jev 可执行集成证据；决策机制提取自实际源码，未经本站性能跑分测试。","claimStatusEn":"Verified with code-level Jev executable integration evidence; decision mechanisms extracted from source.","evidence":[{"url":"https://github.com/teyhouse/jev-secret-detection/blob/fc707416f00a6891b6a03458cc5c6de598fe47b2/README.md","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/teyhouse/jev-secret-detection/blob/fc707416f00a6891b6a03458cc5c6de598fe47b2/main.py","note":"固定版本的 Jev 集成与说明来源"}]},{"id":"noplan-inc:limpet","name":"limpet","author":"noplan-inc","url":"https://github.com/noplan-inc/limpet","category":"SDK & Decision Frameworks","plainSummary":"编程智能体专用 Stop Hook 门禁：防止 Coding Agent 过早宣布完工，用 Jev 依据自然语言规则客观裁定完成度。","plainSummaryEn":"A Stop hook guardrail for coding agents: prevents premature completion by judging plain-language rules via Jev.","jevDecisionPoint":"在 Agent 触发终止指令时拦截，由 Jev 核对任务提示词与上下文是否满足预定验收条件。","jevDecisionPointEn":"Intercepts agent termination signals, using Jev to verify if context meets natural language acceptance criteria.","highlightBenefit":"用轻量级规则裁判阻止代理偷懒或伪造完成，显著提高自动化开发交付完整度。","highlightBenefitEn":"Prevents coding agents from cutting corners or faking completion, boosting autonomous task completion rates.","tags":["multilanguage-sdk","coding-agents"],"stars":2,"forks":1,"openIssues":2,"license":"MIT","lastCommitAt":"2026-09-17T07:53:36Z","createdAt":"2026-09-17T06:37:04Z","metadataFetchedAt":"2026-09-21T22:09:48.075Z","avatarUrl":"https://avatars.githubusercontent.com/u/58355348?v=4","summarySource":"curated","claimStatus":"经由公共仓库代码审查确认真实 Jev 可执行集成证据；决策机制提取自实际源码，未经本站性能跑分测试。","claimStatusEn":"Verified with code-level Jev executable integration evidence; decision mechanisms extracted from source.","evidence":[{"url":"https://github.com/noplan-inc/limpet/blob/7a841b006900cb48df74ad68167dbed17afc504f/README.md","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/noplan-inc/limpet/blob/7a841b006900cb48df74ad68167dbed17afc504f/limpet.py","note":"固定版本的 Jev 集成与说明来源"}]},{"id":"mejiasd3v:pi-jev-router","name":"pi-jev-router","author":"mejiasd3v","url":"https://github.com/mejiasd3v/pi-jev-router","category":"Routing & Cost Optimization","plainSummary":"面向 Pi 编程助手的自动模型路由器：通过 Vercel AI Gateway 集成 Jev，自动为不同编码任务分配合适模型。","plainSummaryEn":"Automatic model router for Pi coding assistant: integrates Jev via Vercel AI Gateway to dispatch tasks efficiently.","jevDecisionPoint":"根据当前代码上下文与用户提问快速评估难度，动态选择小模型或前沿模型处理。","jevDecisionPointEn":"Quickly estimates task complexity from code context and user query, routing between SLMs and frontier models.","highlightBenefit":"为 Pi 终端工具链提供开箱即用的智能化降本分流机制。","highlightBenefitEn":"Brings out-of-the-box intelligent cost-saving routing directly into the Pi terminal workflow.","tags":["llm-routing-cost","typed-decisions"],"stars":11,"forks":1,"openIssues":1,"license":"MIT","lastCommitAt":"2026-09-21T22:08:25Z","createdAt":"2026-09-17T09:24:10Z","metadataFetchedAt":"2026-09-21T22:09:51.072Z","avatarUrl":"https://avatars.githubusercontent.com/u/630925?v=4","summarySource":"curated","claimStatus":"经由公共仓库代码审查确认真实 Jev 可执行集成证据；决策机制提取自实际源码，未经本站性能跑分测试。","claimStatusEn":"Verified with code-level Jev executable integration evidence; decision mechanisms extracted from source.","evidence":[{"url":"https://github.com/mejiasd3v/pi-jev-router/blob/24043d08b4c522a0c07aa432e82cd29bc4e88dd4/README.md","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/mejiasd3v/pi-jev-router/blob/24043d08b4c522a0c07aa432e82cd29bc4e88dd4/index.test.mjs","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/mejiasd3v/pi-jev-router/blob/24043d08b4c522a0c07aa432e82cd29bc4e88dd4/index.ts","note":"固定版本的 Jev 集成与说明来源"}]},{"id":"legacybridge-tech:pi-typesafe-jev","name":"pi-typesafe-jev","author":"legacybridge-tech","url":"https://github.com/legacybridge-tech/pi-typesafe-jev","category":"SDK & Decision Frameworks","plainSummary":"为 Pi 扩展注入 TypeSafe Jev 的 5 种窄域判断工具：把决策权交给模型的同时将动作阈值保留给宿主应用。","plainSummaryEn":"A Pi extension exposing TypeSafe judgments as five narrow tools, keeping threshold and action control in host code.","jevDecisionPoint":"将选择、评分与是非等五类标准判定封装为 Pi 工具，供主代理在需要结构化裁决时调用。","jevDecisionPointEn":"Packages choice, score, and boolean judgments into five Pi tools for structured agent decisions.","highlightBenefit":"严格分离“语义评估”与“业务执行”，确保自动化脚本始终处于受控逻辑内。","highlightBenefitEn":"Strictly decouples semantic evaluation from execution, keeping automated scripts safely constrained.","tags":["multilanguage-sdk","typed-decisions"],"stars":2,"forks":0,"openIssues":0,"license":null,"lastCommitAt":"2026-09-17T04:04:56Z","createdAt":"2026-09-17T03:21:12Z","metadataFetchedAt":"2026-09-21T22:09:54.075Z","avatarUrl":"https://avatars.githubusercontent.com/u/253618205?v=4","summarySource":"curated","claimStatus":"经由公共仓库代码审查确认真实 Jev 可执行集成证据；决策机制提取自实际源码，未经本站性能跑分测试。","claimStatusEn":"Verified with code-level Jev executable integration evidence; decision mechanisms extracted from source.","evidence":[{"url":"https://github.com/legacybridge-tech/pi-typesafe-jev/blob/17e66fcd6ecfbb4c3083f748af510bf59cda985f/README.md","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/legacybridge-tech/pi-typesafe-jev/blob/17e66fcd6ecfbb4c3083f748af510bf59cda985f/src/client.ts","note":"固定版本的 Jev 集成与说明来源"}]},{"id":"grmkris:robo-harness","name":"robo-harness","author":"grmkris","url":"https://github.com/grmkris/robo-harness","category":"Domain & Vertical Tools","plainSummary":"SO-101 机械臂具身智能控制工作台：整合 Bun/Effect 与 Python 驱动，利用 Jev 做关节动作边界决策与预算控制。","plainSummaryEn":"SO-101 robot-arm workbench combining Bun/Effect and Python drivers, using Jev for joint action constraints.","jevDecisionPoint":"在每一步机器人运动规划中，评估空间坐标与传感器状态，由 Jev 在候选安全区间选择执行步骤。","jevDecisionPointEn":"Assesses spatial coordinates and sensor states, using Jev to select safe step actions within bounded budgets.","highlightBenefit":"将 TypeSafe Jev 的低时延离散决策优势拓展至物理世界机械臂实时控制场景。","highlightBenefitEn":"Extends TypeSafe Jev low-latency discrete judgments to physical-world robotic arm motion control.","tags":["domain-workflows","typed-decisions"],"stars":1,"forks":0,"openIssues":0,"license":null,"lastCommitAt":"2026-09-21T21:31:39Z","createdAt":"2026-09-07T07:53:41Z","metadataFetchedAt":"2026-09-21T22:09:57.036Z","avatarUrl":"https://avatars.githubusercontent.com/u/15545195?v=4","summarySource":"curated","claimStatus":"经由公共仓库代码审查确认真实 Jev 可执行集成证据；决策机制提取自实际源码，未经本站性能跑分测试。","claimStatusEn":"Verified with code-level Jev executable integration evidence; decision mechanisms extracted from source.","evidence":[{"url":"https://github.com/grmkris/robo-harness/blob/143699975b09df94b4d986329cdae5fdf9541d31/README.md","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/grmkris/robo-harness/blob/143699975b09df94b4d986329cdae5fdf9541d31/apps/server/src/decision/jev.test.ts","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/grmkris/robo-harness/blob/143699975b09df94b4d986329cdae5fdf9541d31/apps/server/src/decision/jev.ts","note":"固定版本的 Jev 集成与说明来源"}]},{"id":"siroccomask:snake-jev","name":"snake-jev","author":"siroccomask","url":"https://github.com/siroccomask/snake-jev","category":"High-Frequency & Simulation","plainSummary":"Jev 并行概率驱动的贪吃蛇实时游戏控制：在每个游戏 Tick 仅需一次并发 API 调用即可选出最佳转向。","plainSummaryEn":"Real-time Snake game driven by parallel Jev assessments, deciding optimal turns in a single API call per tick.","jevDecisionPoint":"实时扫描蛇头四周障碍与食物方位，由 Jev 同时对上下左右四个方向输出存活与逼近概率。","jevDecisionPointEn":"Scans obstacles and food locations per tick, querying Jev in parallel for survival and approach probabilities.","highlightBenefit":"验证了 Jev 在高频、确定性帧率要求下的并发低时延决策表现。","highlightBenefitEn":"Demonstrates Jev high-frequency, deterministic low-latency performance in real-time game ticks.","tags":["games-simulation","typed-decisions"],"stars":0,"forks":0,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-17T10:02:07Z","createdAt":"2026-09-17T10:02:13Z","metadataFetchedAt":"2026-09-21T22:10:00.053Z","avatarUrl":"https://avatars.githubusercontent.com/u/9921926?v=4","summarySource":"curated","claimStatus":"经由公共仓库代码审查确认真实 Jev 可执行集成证据；决策机制提取自实际源码，未经本站性能跑分测试。","claimStatusEn":"Verified with code-level Jev executable integration evidence; decision mechanisms extracted from source.","evidence":[{"url":"https://github.com/siroccomask/snake-jev/blob/86f01b686df2e6d5b566b80d015de9b8b34450a8/README.md","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/siroccomask/snake-jev/blob/86f01b686df2e6d5b566b80d015de9b8b34450a8/jev_controller.py","note":"固定版本的 Jev 集成与说明来源"}]},{"id":"mattn:sqlite3-jev","name":"sqlite3-jev","author":"mattn","url":"https://github.com/mattn/sqlite3-jev","category":"Domain & Vertical Tools","plainSummary":"将 TypeSafe Jev 判断能力下沉为 SQLite 自定义 SQL 函数的 C 语言扩展：直接在 SQL 查询中实现语义打分与选择。","plainSummaryEn":"SQLite C extension enabling TypeSafe Jev judgments as native SQL functions for semantic scoring and choices.","jevDecisionPoint":"在 SQLite 执行引擎中注册 jev_choice、jev_score 等自定义函数，在查询扫描行时直接调用 Jev API。","jevDecisionPointEn":"Registers custom functions like jev_choice and jev_score directly inside the SQLite query pipeline.","highlightBenefit":"无需额外编写业务胶水代码，即可在大规模结构化关系型数据上执行行级语义分类与推理。","highlightBenefitEn":"Enables row-level semantic classification directly inside relational SQL queries without glue code.","tags":["domain-workflows","typed-decisions"],"stars":3,"forks":0,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-18T03:12:25Z","createdAt":"2026-09-18T03:12:25Z","metadataFetchedAt":"2026-09-21T22:10:03.084Z","avatarUrl":"https://avatars.githubusercontent.com/u/10111?v=4","summarySource":"curated","claimStatus":"经由公共仓库代码审查确认真实 Jev 可执行集成证据；决策机制提取自实际源码，未经本站性能跑分测试。","claimStatusEn":"Verified with code-level Jev executable integration evidence; decision mechanisms extracted from source.","evidence":[{"url":"https://github.com/mattn/sqlite3-jev/blob/2172600da148182703f8f221089d9304d0fb97f9/README.md","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/mattn/sqlite3-jev/blob/2172600da148182703f8f221089d9304d0fb97f9/jev.c","note":"固定版本的 Jev 集成与说明来源"}]},{"id":"kubet:azdaja","name":"azdaja","author":"kubet","url":"https://github.com/kubet/azdaja","category":"Context GC & Filter","plainSummary":"Azdaja 是与 harness 无关的递归语言模型层，将完整资料保存在本地求值器中并只对选定内容做语义递归，可配合 Jev 进行类型化语义判断。","plainSummaryEn":"Bare, open-source RLM layer for existing coding agents","jevDecisionPoint":"Jev 对选定的源码窗口就显式问题返回Noul/Choice/Score概率，供RLM决定下一步检查、合并或探索的内容。","jevDecisionPointEn":"Jev receives selected source material and explicit questions. It can rank passages, classify records, assess whether evidence supports a claim and judge whether records match for semantic joins. Its returned probabilities let the RLM compare results and decide what to inspect, combine or explore next.","highlightBenefit":"把语义判断接到已有的命令行工作流。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["context-compaction","cli-git-gates","typed-decisions"],"stars":10,"forks":0,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-19T11:00:45Z","createdAt":"2026-08-11T23:07:02Z","metadataFetchedAt":"2026-09-21T22:10:06.057Z","avatarUrl":"https://avatars.githubusercontent.com/u/41026612?v=4","summarySource":"readme-extractive","claimStatus":"优先保留投稿者与仓库原文，缺失语言自动补充；自动检查仅确认 Jev 集成证据，未经本站运行或性能复测。","claimStatusEn":"Author and repository text is preserved where clear; missing languages are enriched automatically. Checks establish source-level Jev integration, not runtime, safety, or performance validation.","evidence":[{"url":"https://github.com/kubet/azdaja/blob/0ac5f0de0c99a5db0d32fa8691dbf08c14663923/README.md","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/kubet/azdaja/blob/0ac5f0de0c99a5db0d32fa8691dbf08c14663923/examples/jev_batch_review.py","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/kubet/azdaja/blob/0ac5f0de0c99a5db0d32fa8691dbf08c14663923/examples/jev_review_queue.py","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/kubet/azdaja/blob/0ac5f0de0c99a5db0d32fa8691dbf08c14663923/src/judge_batch.rs","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/kubet/azdaja/blob/0ac5f0de0c99a5db0d32fa8691dbf08c14663923/src/judge.rs","note":"固定版本的 Jev 集成与说明来源"}]},{"id":"rajdhakad9826:jev-router","name":"jev-router","author":"rajdhakad9826","url":"https://github.com/rajdhakad9826/jev-router","category":"Routing & Cost Optimization","plainSummary":"基于 Jev 评分概率与期望损失最小化（Expected Loss Minimization）实现的成本敏感型大模型路由器。","plainSummaryEn":"Cost-aware LLM router that picks the cheapest model capable of handling a query, using TypeSafe's Jev for fast classification instead of an LLM call.","jevDecisionPoint":"决策说明：\n在 classify 方法中使用 Jev score 对模型能力层级进行概率评分，后续结合成本损失矩阵计算最优模型。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"把语义判断接到已有的命令行工作流。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["llm-routing-cost","typed-decisions"],"stars":8,"forks":0,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-21T09:00:19Z","createdAt":"2026-09-19T14:56:21Z","metadataFetchedAt":"2026-09-21T22:10:09.062Z","avatarUrl":"https://avatars.githubusercontent.com/u/111659873?v=4","summarySource":"readme-extractive","claimStatus":"优先保留投稿者与仓库原文，缺失语言自动补充；自动检查仅确认 Jev 集成证据，未经本站运行或性能复测。","claimStatusEn":"Author and repository text is preserved where clear; missing languages are enriched automatically. Checks establish source-level Jev integration, not runtime, safety, or performance validation.","evidence":[{"url":"https://github.com/rajdhakad9826/jev-router/blob/9c332db56da87a559a1e18d0331c8fcb303b40fd/README.md","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/rajdhakad9826/jev-router/blob/9c332db56da87a559a1e18d0331c8fcb303b40fd/src/router.ts","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/rajdhakad9826/jev-router/blob/9c332db56da87a559a1e18d0331c8fcb303b40fd/src/jev/classifier.ts","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/rajdhakad9826/jev-router/blob/9c332db56da87a559a1e18d0331c8fcb303b40fd/src/core/loss.ts","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/rajdhakad9826/jev-router/blob/9c332db56da87a559a1e18d0331c8fcb303b40fd/src/core/normalise.ts","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/rajdhakad9826/jev-router/blob/9c332db56da87a559a1e18d0331c8fcb303b40fd/src/core/selection.ts","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/rajdhakad9826/jev-router/blob/9c332db56da87a559a1e18d0331c8fcb303b40fd/src/index.ts","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/rajdhakad9826/jev-router/blob/9c332db56da87a559a1e18d0331c8fcb303b40fd/src/types.ts","note":"固定版本的 Jev 集成与说明来源"}]},{"id":"logicrw:ask-jev","name":"ask-jev","author":"logicrw","url":"https://github.com/logicrw/ask-jev","category":"CLI & Pipelines","plainSummary":"面向 AI 编程智能体与 CLI 管道的有界决策与原文提纯工具，硬性 280ms 时限，纯 Python 标准库零依赖，全链路 Fail-Open 优雅降级。","plainSummaryEn":"Ultra-fast, fail-open advisory decisions and verbatim extractive reading view for AI coding agents and CLI pipelines","jevDecisionPoint":"决策说明：\n在 ask_decision 与 auto_mark_text 中调用 Jev 的 Choice 与 Noul 原语，执行快速语义裁决与物理行保真提纯。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"把语义判断接到已有的命令行工作流。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["cli-git-gates","coding-agents","context-compaction"],"stars":1,"forks":0,"openIssues":0,"license":"GPL-3.0","lastCommitAt":"2026-09-21T09:08:12Z","createdAt":"2026-09-20T11:23:53Z","metadataFetchedAt":"2026-09-21T22:10:12.091Z","avatarUrl":"https://avatars.githubusercontent.com/u/34853455?v=4","summarySource":"readme-extractive","claimStatus":"优先保留投稿者与仓库原文，缺失语言自动补充；自动检查仅确认 Jev 集成证据，未经本站运行或性能复测。","claimStatusEn":"Author and repository text is preserved where clear; missing languages are enriched automatically. Checks establish source-level Jev integration, not runtime, safety, or performance validation.","evidence":[{"url":"https://github.com/logicrw/ask-jev/blob/8485cf82dc2ed5e8b0950933384021fc116b00b6/README.md","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/logicrw/ask-jev/blob/8485cf82dc2ed5e8b0950933384021fc116b00b6/scripts/ask_jev.py","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/logicrw/ask-jev/blob/8485cf82dc2ed5e8b0950933384021fc116b00b6/scripts/jev_context.py","note":"固定版本的 Jev 集成与说明来源"}]},{"id":"lukstei:slop-grader","name":"slop-grader","author":"lukstei","url":"https://github.com/lukstei/slop-grader","category":"CLI & Pipelines","plainSummary":"这是一个基于规则的 CLI 工具，用 Jev 的 Score 做文档级打分、用 Choice 和 Noul 做逐行违规标记，并把结果交给 Agent 去修复文本和 Markdown 文件。","plainSummaryEn":"Rule-based CLI and agent skill that evaluates text and markdown files against custom rulesets for AI slop, grammar, and technical doc quality using Jev scores and line-by-line violation flags, then guides an AI agent to auto-fix violations.","jevDecisionPoint":"决策说明：\n在 `src/providers/jev.ts` 的 `evaluateBatch` 中调用 `@typesafe-ai/sdk` (`client.system_one`)，对文档级别使用 `score` 进行标准维度评分，对行级别使用 `choice` / `noul` 进行细粒度违规检测，并将结构化违规结果输出给下游 AI agent 自动化修复。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"增加一组可记录、可对照的判断信号。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["cli-git-gates","coding-agents","typed-decisions"],"stars":9,"forks":1,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-21T22:05:25Z","createdAt":"2026-09-20T07:12:51Z","metadataFetchedAt":"2026-09-21T22:10:15.035Z","avatarUrl":"https://avatars.githubusercontent.com/u/292363?v=4","summarySource":"readme-extractive","claimStatus":"优先保留投稿者与仓库原文，缺失语言自动补充；自动检查仅确认 Jev 集成证据，未经本站运行或性能复测。","claimStatusEn":"Author and repository text is preserved where clear; missing languages are enriched automatically. Checks establish source-level Jev integration, not runtime, safety, or performance validation.","evidence":[{"url":"https://github.com/lukstei/slop-grader/blob/aedb334961fd201d958a7dbfc3a0ba09c8c8c99d/README.md","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/lukstei/slop-grader/blob/aedb334961fd201d958a7dbfc3a0ba09c8c8c99d/src/providers/jev.test.ts","note":"固定版本的 Jev 集成与说明来源"}]},{"id":"lychoon:paper-radar-jev","name":"paper-radar-jev","author":"LYchoon","url":"https://github.com/LYchoon/paper-radar-jev","category":"CLI & Pipelines","plainSummary":"以上指令會保留已存在的本機設定。編輯專案根目錄的 `.env`，將空白值換成自己的 key：","plainSummaryEn":"An automated research paper radar that fetches the latest papers from arXiv, evaluates their relevance to a configurable research profile using TypeSafe AI, and ranks them by relevance score. Designed for personalized, daily literature discovery across different research domains.","jevDecisionPoint":"决策说明： 在 src/paper_radar/evaluator.py 的 evaluate_paper() 中，通过 client.system_one(...) 将论文信息和 research profile 发送给 TypeSafe System One，并从返回的 response.nouls[...] 中取得概率结果作为 relevance_score。这个位置就是项目实际进行论文相关性决策的地方。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"把语义判断接到已有的命令行工作流。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["cli-git-gates","classification-ranking","search-retrieval"],"stars":0,"forks":0,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-20T04:08:37Z","createdAt":"2026-09-20T03:11:13Z","metadataFetchedAt":"2026-09-21T22:10:18.061Z","avatarUrl":"https://avatars.githubusercontent.com/u/158578641?v=4","summarySource":"readme-extractive","claimStatus":"优先保留投稿者与仓库原文，缺失语言自动补充；自动检查仅确认 Jev 集成证据，未经本站运行或性能复测。","claimStatusEn":"Author and repository text is preserved where clear; missing languages are enriched automatically. Checks establish source-level Jev integration, not runtime, safety, or performance validation.","evidence":[{"url":"https://github.com/LYchoon/paper-radar-jev/blob/fbadf01c69806730eb157bdc97f662a38b522593/README.md","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/LYchoon/paper-radar-jev/blob/fbadf01c69806730eb157bdc97f662a38b522593/README.zh-TW.md","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/LYchoon/paper-radar-jev/blob/fbadf01c69806730eb157bdc97f662a38b522593/src/paper_radar/models.py","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/LYchoon/paper-radar-jev/blob/fbadf01c69806730eb157bdc97f662a38b522593/main.py","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/LYchoon/paper-radar-jev/blob/fbadf01c69806730eb157bdc97f662a38b522593/src/paper_radar/__init__.py","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/LYchoon/paper-radar-jev/blob/fbadf01c69806730eb157bdc97f662a38b522593/src/paper_radar/__main__.py","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/LYchoon/paper-radar-jev/blob/fbadf01c69806730eb157bdc97f662a38b522593/src/paper_radar/cli.py","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/LYchoon/paper-radar-jev/blob/fbadf01c69806730eb157bdc97f662a38b522593/src/paper_radar/config.py","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/LYchoon/paper-radar-jev/blob/fbadf01c69806730eb157bdc97f662a38b522593/src/paper_radar/evaluator.py","note":"固定版本的 Jev 集成与说明来源"}]},{"id":"patrick-sch03:jev-issue-radar","name":"jev-issue-radar","author":"Patrick-SCH03","url":"https://github.com/Patrick-SCH03/jev-issue-radar","category":"Domain & Vertical Tools","plainSummary":"Jev Issue Radar 是一个只读的 GitHub 重复问题分类看板，它检索候选问题并用 Jev 的 Choice 将每对问题判为重复、相关、不同或信息不足，同时并排展示双方原始报告中的选中段落供维护者审查。","plainSummaryEn":"Jev Issue Radar is a read-only dashboard for GitHub duplicate-issue triage. It retrieves likely candidates, asks Jev whether each pair is duplicate, related, distinct, or insufficiently documented, and shows selected passages from both original reports for a maintainer to review.","jevDecisionPoint":"对每对原始问题与候选问题判定重复、相关、不同或信息不足，并选择原因类别与双方原文证据行ID。","jevDecisionPointEn":"The app never closes issues, posts comments, or changes labels. Retrieval is currently lexical and bounded to a recent-item scan; this limitation is explicit in the UI and README.","highlightBenefit":"把语义判断接到已有的命令行工作流。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["domain-workflows","search-retrieval","classification-ranking"],"stars":1,"forks":0,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-20T15:06:30Z","createdAt":"2026-09-20T14:10:07Z","metadataFetchedAt":"2026-09-21T22:10:21.129Z","avatarUrl":"https://avatars.githubusercontent.com/u/191001608?v=4","summarySource":"readme-extractive","claimStatus":"优先保留投稿者与仓库原文，缺失语言自动补充；自动检查仅确认 Jev 集成证据，未经本站运行或性能复测。","claimStatusEn":"Author and repository text is preserved where clear; missing languages are enriched automatically. Checks establish source-level Jev integration, not runtime, safety, or performance validation.","evidence":[{"url":"https://github.com/Patrick-SCH03/jev-issue-radar/blob/6ea5391aa229ce77466774d2ee6af1abb631bb7a/README.md","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/Patrick-SCH03/jev-issue-radar/blob/6ea5391aa229ce77466774d2ee6af1abb631bb7a/lib/jev.mjs","note":"固定版本的 Jev 集成与说明来源"}]},{"id":"spoonnotfound:soupbase","name":"soupbase","author":"spoonnotfound","url":"https://github.com/spoonnotfound/soupbase","category":"High-Frequency & Simulation","plainSummary":"汤底 Soupbase 是中英文海龟汤游戏，由 Jev 判定玩家提问和还原内容；应用代码校验结构化结果与置信度，决定是否通关。支持私有创作、链接分享和自行部署。","plainSummaryEn":"Soupbase is a bilingual Chinese-English Turtle Soup game where Jev judges player questions and reconstructions, and the app checks structured Choice results and confidence to decide clearance.","jevDecisionPoint":"两类判断均使用 Choice；没有使用 Score 或 Noul。提示与主动揭晓读取已保存内容，不依赖模型生成。这里选择游戏分类，不主张高频性能，也不声明未经测量的速度、成本或准确率。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"把语义判断接进已有的数据查询流程。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["games-simulation","typed-decisions"],"stars":3,"forks":0,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-21T16:21:24Z","createdAt":"2026-09-19T06:50:42Z","metadataFetchedAt":"2026-09-21T22:10:24.079Z","avatarUrl":"https://avatars.githubusercontent.com/u/20537546?v=4","summarySource":"readme-extractive","claimStatus":"优先保留投稿者与仓库原文，缺失语言自动补充；自动检查仅确认 Jev 集成证据，未经本站运行或性能复测。","claimStatusEn":"Author and repository text is preserved where clear; missing languages are enriched automatically. Checks establish source-level Jev integration, not runtime, safety, or performance validation.","evidence":[{"url":"https://github.com/spoonnotfound/soupbase/blob/4e071a5a500f7db2a8ec4f5a3f99aa02a4c784aa/README.md","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/spoonnotfound/soupbase/blob/4e071a5a500f7db2a8ec4f5a3f99aa02a4c784aa/src/app/api/%5B...path%5D/route.ts","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/spoonnotfound/soupbase/blob/4e071a5a500f7db2a8ec4f5a3f99aa02a4c784aa/src/components/model-trace.tsx","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/spoonnotfound/soupbase/blob/4e071a5a500f7db2a8ec4f5a3f99aa02a4c784aa/src/server/api.ts","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/spoonnotfound/soupbase/blob/4e071a5a500f7db2a8ec4f5a3f99aa02a4c784aa/src/server/model.ts","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/spoonnotfound/soupbase/blob/4e071a5a500f7db2a8ec4f5a3f99aa02a4c784aa/src/app/%5Blocale%5D/page.tsx","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/spoonnotfound/soupbase/blob/4e071a5a500f7db2a8ec4f5a3f99aa02a4c784aa/src/app/layout.tsx","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/spoonnotfound/soupbase/blob/4e071a5a500f7db2a8ec4f5a3f99aa02a4c784aa/src/app/page.tsx","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/spoonnotfound/soupbase/blob/4e071a5a500f7db2a8ec4f5a3f99aa02a4c784aa/src/components/app.tsx","note":"固定版本的 Jev 集成与说明来源"}]},{"id":"keltokhy:jselect","name":"jselect","author":"keltokhy","url":"https://github.com/keltokhy/jselect","category":"Data & Search","plainSummary":"jselect 根据任务挑选带来源引用的原文片段，并用 Jev 判断相关性，在 Token 预算内组装可直接供 Agent 阅读的上下文。","plainSummaryEn":"Selects source-linked evidence within a token budget using Jev Noul relevance judgments and local diversity-aware selection.","jevDecisionPoint":"Jev 对每个文本片段做出 Noul 二元判断，决定其是否为当前任务的有用证据。","jevDecisionPointEn":"Jev judges whether each passage is useful evidence for the supplied task; local code selects verbatim passages and citations within the token budget.","highlightBenefit":"增加一组可记录、可对照的判断信号。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["search-retrieval","context-compaction","typed-decisions"],"stars":3,"forks":0,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-21T03:10:25Z","createdAt":"2026-09-19T19:28:43Z","metadataFetchedAt":"2026-09-21T22:10:27.050Z","avatarUrl":"https://avatars.githubusercontent.com/u/83777601?v=4","summarySource":"readme-extractive","claimStatus":"优先保留投稿者与仓库原文，缺失语言自动补充；自动检查仅确认 Jev 集成证据，未经本站运行或性能复测。","claimStatusEn":"Author and repository text is preserved where clear; missing languages are enriched automatically. Checks establish source-level Jev integration, not runtime, safety, or performance validation.","evidence":[{"url":"https://github.com/keltokhy/jselect/blob/460b217110a6fa2f3546813f1dc4f77b062e82cd/README.md","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/keltokhy/jselect/blob/460b217110a6fa2f3546813f1dc4f77b062e82cd/src/jselect/judge.py","note":"固定版本的 Jev 集成与说明来源"}]},{"id":"keltokhy:jgrep","name":"jgrep","author":"keltokhy","url":"https://github.com/keltokhy/jgrep","category":"CLI & Pipelines","plainSummary":"jgrep 是用自然语言描述替代正则表达式来过滤文本行、结构化记录、函数和 diff 片段的命令行工具，每条记录交由 Jev 作 Noul 判断后输出符合项。","plainSummaryEn":"Filters text, structured records, functions, and diff hunks against plain-English descriptions using Jev Noul judgments.","jevDecisionPoint":"Jev 对每个输入单元（行、段落、函数、diff hunk）判断是否符合用户自然语言描述，返回概率后由本地阈值过滤。","jevDecisionPointEn":"Jev judges whether each input unit matches the user description; local code applies the probability threshold and returns matching source material.","highlightBenefit":"把语义判断接到已有的命令行工作流。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["cli-git-gates","search-retrieval","typed-decisions"],"stars":16,"forks":2,"openIssues":2,"license":"MIT","lastCommitAt":"2026-09-21T03:10:25Z","createdAt":"2026-09-18T11:11:25Z","metadataFetchedAt":"2026-09-21T22:10:30.062Z","avatarUrl":"https://avatars.githubusercontent.com/u/83777601?v=4","summarySource":"readme-extractive","claimStatus":"优先保留投稿者与仓库原文，缺失语言自动补充；自动检查仅确认 Jev 集成证据，未经本站运行或性能复测。","claimStatusEn":"Author and repository text is preserved where clear; missing languages are enriched automatically. Checks establish source-level Jev integration, not runtime, safety, or performance validation.","evidence":[{"url":"https://github.com/keltokhy/jgrep/blob/fdceb6bdf79165a133667b8e57f3b7244545f0a2/README.md","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/keltokhy/jgrep/blob/fdceb6bdf79165a133667b8e57f3b7244545f0a2/src/jgrep/__init__.py","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/keltokhy/jgrep/blob/fdceb6bdf79165a133667b8e57f3b7244545f0a2/src/jgrep/__main__.py","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/keltokhy/jgrep/blob/fdceb6bdf79165a133667b8e57f3b7244545f0a2/src/jgrep/cli.py","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/keltokhy/jgrep/blob/fdceb6bdf79165a133667b8e57f3b7244545f0a2/src/jgrep/code_inputs.py","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/keltokhy/jgrep/blob/fdceb6bdf79165a133667b8e57f3b7244545f0a2/src/jgrep/core.py","note":"固定版本的 Jev 集成与说明来源"}]},{"id":"ctmx:openrouter-jev-mcp","name":"openrouter-jev-mcp","author":"ctmx","url":"https://github.com/ctmx/openrouter-jev-mcp","category":"MCP & Integrations","plainSummary":"该项目是通过 OpenRouter 提供 Jev Choice、Score 和 Noul 类型化决策的 Python 网关和 MCP 服务器，供编码 Agent 调用。","plainSummaryEn":"A Python decision gateway and Model Context Protocol (MCP) server exposing TypeSafe's Jev model through OpenRouter's decisions endpoint to Claude Code, Codex, and Cursor agents.","jevDecisionPoint":"Jev 对代码 diff 状态执行根因多选一分类、安全风险连续打分和是否可上线的二元门禁判断。","jevDecisionPointEn":"The server exposes Jev Choice (`jev_classify`), Score (`jev_score`), and Noul (`jev_check`) decisions over standard stdio MCP to AI coding agents, executing requests against OpenRouter's decisions endpoint:","highlightBenefit":"让现有 Agent 通过通用接口使用 Jev。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["mcp-integrations","typed-decisions"],"stars":0,"forks":0,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-20T18:47:18Z","createdAt":"2026-09-19T18:05:20Z","metadataFetchedAt":"2026-09-21T22:10:33.079Z","avatarUrl":"https://avatars.githubusercontent.com/u/49662053?v=4","summarySource":"readme-extractive","claimStatus":"优先保留投稿者与仓库原文，缺失语言自动补充；自动检查仅确认 Jev 集成证据，未经本站运行或性能复测。","claimStatusEn":"Author and repository text is preserved where clear; missing languages are enriched automatically. Checks establish source-level Jev integration, not runtime, safety, or performance validation.","evidence":[{"url":"https://github.com/ctmx/openrouter-jev-mcp/blob/e9d4bb38dbba7c7dba0cd4d0504ada741c87d1d7/README.md","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/ctmx/openrouter-jev-mcp/blob/e9d4bb38dbba7c7dba0cd4d0504ada741c87d1d7/examples/demo_typesafe_sdk.py","note":"固定版本的 Jev 集成与说明来源"}]},{"id":"sutro-sh:jev-align","name":"jev-align","author":"sutro-sh","url":"https://github.com/sutro-sh/jev-align","category":"CLI & Pipelines","plainSummary":"jev-align 是一个实验性 CLI，用 Jev 构建 AI Functions，通过挑选不确定样本请用户标注并用 GEPA 改进函数定义。","plainSummaryEn":"1. Evaluates the configured dataset and measures uncertainty.\n2. Selects ambiguous rows plus a random audit sample for you to label.\n3. Uses your accumulated labels and optional rationales to run GEPA.\n4. Shows the score, certainty change, and proposed definition diff.\n5. Lets you accept, reject, rewind, or resume later.","jevDecisionPoint":"Jev 对每条数据行执行二分类/多分类/多标签/等级评分判断并返回校准概率与置信度，用于筛选不确定样本和评估候选定义。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"增加一组可记录、可对照的判断信号。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["cli-git-gates","typed-decisions"],"stars":264,"forks":20,"openIssues":1,"license":"Apache-2.0","lastCommitAt":"2026-09-20T20:49:51Z","createdAt":"2026-09-19T02:12:23Z","metadataFetchedAt":"2026-09-21T22:10:36.059Z","avatarUrl":"https://avatars.githubusercontent.com/u/230064936?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/sutro-sh/jev-align/blob/3d997fc76593036655c28d4964de43b55f81fe2c/src/jev_align/jev.py","note":"自动发现的 Jev 集成证据"}]},{"id":"savka777:jev-use","name":"jev-use","author":"savka777","url":"https://github.com/savka777/jev-use","category":"Browser & OS Action","plainSummary":"这是在 macOS 上用语音或文字下指令、由 Jev 读取 Accessibility 树并操作屏幕元素的电脑使用工具。","plainSummaryEn":"Voice and typed computer use for macOS. You say what you want. Jev picks the next on-screen action. macOS performs it. No screenshots: the app reads the screen through the Accessibility tree.","jevDecisionPoint":"Jev 根据用户指令、当前应用窗口和编号后的可操作目标列表，选择下一步应执行的操作及其目标控件。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"让界面选择与执行分开，便于检查每一步。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["browser-automation","typed-decisions"],"stars":84,"forks":10,"openIssues":2,"license":"MIT","lastCommitAt":"2026-09-19T16:26:24Z","createdAt":"2026-09-19T14:05:35Z","metadataFetchedAt":"2026-09-21T22:10:39.046Z","avatarUrl":"https://avatars.githubusercontent.com/u/103605876?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/savka777/jev-use/blob/8907f85354addfa3d2b78f6a462087e4c73310b8/Sources/JevCore/Decision.swift","note":"自动发现的 Jev 集成证据"}]},{"id":"bodila51:grok-bot-jev","name":"grok-bot-jev","author":"Bodila51","url":"https://github.com/Bodila51/grok-bot-jev","category":"Routing & Cost Optimization","plainSummary":"该项目为 Grok Bot 提供一个 Python 路由器，它在执行高成本操作前调用 Jev 对请求分类并返回明确的路由动作，同时附带可供 Grok Bot 技能使用的模板和示例。","plainSummaryEn":"Connect TypeSafe Jev to Grok Bot as a cheap decision layer - usage gates, skill template, examples","jevDecisionPoint":"Jev 在执行昂贵的研究、浏览器、重试或子代理工作之前，决定执行 reuse_cache、stop_retry、research_capped 等路由动作。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"把语义判断接到已有的命令行工作流。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["llm-routing-cost","typed-decisions"],"stars":70,"forks":7,"openIssues":2,"license":"MIT","lastCommitAt":"2026-09-20T00:13:30Z","createdAt":"2026-09-19T17:15:48Z","metadataFetchedAt":"2026-09-21T22:10:42.117Z","avatarUrl":"https://avatars.githubusercontent.com/u/191466064?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/Bodila51/grok-bot-jev/blob/1583e09928c138aeac0aa89818c67ea41f08e807/src/jev_client.py","note":"自动发现的 Jev 集成证据"}]},{"id":"devin-axis:jev-dsh-decision","name":"jev-dsh-decision","author":"Devin-AXIS","url":"https://github.com/Devin-AXIS/jev-dsh-decision","category":"SDK & Decision Frameworks","plainSummary":"Jev DSH 决策引擎｜面向 Agent Harness 的结构化决策插件。原生支持 DeepSeek Harness，通过 iPolloWork 支持 OpenCode、Codex Harness。","plainSummaryEn":"Provides Jev structured decision support for Agent Harness to recommend tools, Skills and Agents and return judgments with probabilities, with a native DeepSeek Harness plugin and an iPolloWork entry serving OpenCode, DeepSeek Harness and Codex Harness.","jevDecisionPoint":"Jev 根据任务在当前可用工具、Skill 和 Agent 之间选择最合适项并对产出质量进行结构化评分。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"把选择和打分接进现有程序；暂无可核验的性能对照。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["multilanguage-sdk","typed-decisions"],"stars":65,"forks":25,"openIssues":1,"license":null,"lastCommitAt":"2026-09-20T04:59:27Z","createdAt":"2026-09-20T04:39:05Z","metadataFetchedAt":"2026-09-21T22:10:45.045Z","avatarUrl":"https://avatars.githubusercontent.com/u/228423313?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/Devin-AXIS/jev-dsh-decision/blob/adc88caa9bf174d367e79a5f254c7936bcef088e/service/jev.mjs","note":"自动发现的 Jev 集成证据"}]},{"id":"samdotmak:jev-recall","name":"jev-recall","author":"samdotmak","url":"https://github.com/samdotmak/jev-recall","category":"Context GC & Filter","plainSummary":"Jev Recall 针对用户请求，用 Jev 模型对每条记忆逐一判断是否相关并筛选出相关记忆。","plainSummaryEn":"Retrieve by relevance, not resemblance: filter an AI assistant's memories with TypeSafe's Jev","jevDecisionPoint":"对每条记忆判断它是否与当前情景相关、是否会改变助手的行为。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"把语义判断接进已有的数据查询流程。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["context-compaction","typed-decisions"],"stars":31,"forks":1,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-20T00:17:14Z","createdAt":"2026-09-19T20:08:14Z","metadataFetchedAt":"2026-09-21T22:10:48.086Z","avatarUrl":"https://avatars.githubusercontent.com/u/110200324?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/samdotmak/jev-recall/blob/d3e4acfb45c7cf6231f1621b7dafb09d21e95862/src/jev_recall/core.py","note":"自动发现的 Jev 集成证据"}]},{"id":"fstandhartinger:jevbench","name":"jevbench","author":"fstandhartinger","url":"https://github.com/fstandhartinger/jevbench","category":"SDK & Decision Frameworks","plainSummary":"该仓库提供 JevBench 基准，向 Jev 类决策模型输入状态和有界评分规则，并对返回的类型化答案按 Intelligence、Calibration、Speed 和 Cost 四个维度计分。","plainSummaryEn":"JevBench v1 - a benchmark for Jev-class typed decision models: smart, cheap, fast, reliable, open.","jevDecisionPoint":"Jev 根据输入的 state 和有界 rubric，对 noul/choice/score 三类问题返回类型化答案及每个选项的概率分布。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"把选择和打分接进现有程序；暂无可核验的性能对照。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["multilanguage-sdk","evaluation-benchmarks","typed-decisions"],"stars":57,"forks":3,"openIssues":20,"license":"MIT","lastCommitAt":"2026-09-21T20:43:51Z","createdAt":"2026-09-19T07:07:36Z","metadataFetchedAt":"2026-09-21T22:10:51.104Z","avatarUrl":"https://avatars.githubusercontent.com/u/1905188?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/fstandhartinger/jevbench/blob/275763201a29d6083d4ee1431d709c296ef81281/jevbench/adapters/typesafe.py","note":"自动发现的 Jev 集成证据"}]},{"id":"zaious:jev-capability-atlas","name":"jev-capability-atlas","author":"Zaious","url":"https://github.com/Zaious/jev-capability-atlas","category":"SDK & Decision Frameworks","plainSummary":"**機制上**：它很快、很便宜，只能做「選一個選項/打個分/回答是非」這種窄判斷，不會寫文字解釋自己在想什麼。因為答案空間是你自己先定義好的，它**結構上不可能吐出選項清單以外的東西**——這跟自由生成文字的模型偶爾格式跑掉、甚至生出一個你沒列的分類，是不同等級的保證，不是機率低，是型別上不可能（但這只保證答案落在清單裡，不保證選到的那個是對的，細節見下面「不是瞎猜」那節）。","plainSummaryEn":"This repository collects real Jev API-call receipts, test suites, and bilingual guides to map which narrow-decision tasks suit Jev and how Agents should evaluate and report fit.","jevDecisionPoint":"Jev 在各测试套件中执行窄判断，如引用支持度、反讽识别、历史问答等选项选择与是非打分决策。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"把选择和打分接进现有程序；暂无可核验的性能对照。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["multilanguage-sdk","typed-decisions"],"stars":23,"forks":5,"openIssues":3,"license":"MIT","lastCommitAt":"2026-09-21T17:01:19Z","createdAt":"2026-09-18T21:30:40Z","metadataFetchedAt":"2026-09-21T22:10:54.083Z","avatarUrl":"https://avatars.githubusercontent.com/u/128442444?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/Zaious/jev-capability-atlas/blob/b85bae4925f77a8d9576f63882a94521936d3c22/scripts/common/jev_client.py","note":"自动发现的 Jev 集成证据"}]},{"id":"nassim-arifette:jevgrep","name":"jevgrep","author":"nassim-arifette","url":"https://github.com/nassim-arifette/jevgrep","category":"Codebase & Graph Pathfinding","plainSummary":"JevGrep 是一个基于 Jev 的语义代码搜索工具，可通过 CLI 或 MCP 按行为描述查找代码，并返回带文件路径和行号的原文片段。","plainSummaryEn":"Jev-powered semantic code search for coding agents — find behavior across repositories via CLI or MCP, with exact source excerpts and line numbers.","jevDecisionPoint":"Jev 对每个代码片段是否语义匹配自然语言查询进行 Noul 评分排序，以决定返回哪些源码摘录。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"让现有 Agent 通过通用接口使用 Jev。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["ast-code-graph","cli-git-gates","mcp-integrations"],"stars":44,"forks":5,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-20T17:01:42Z","createdAt":"2026-09-20T00:27:29Z","metadataFetchedAt":"2026-09-21T22:10:57.052Z","avatarUrl":"https://avatars.githubusercontent.com/u/76560658?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/nassim-arifette/jevgrep/blob/52e2d895b69d3bd41138b5f9091d7eed3e2cf4a0/experiments/jev-contract/probe.mjs","note":"自动发现的 Jev 集成证据"}]},{"id":"khordoo:jev-reflex-autonomy-lab","name":"jev-reflex-autonomy-lab","author":"khordoo","url":"https://github.com/khordoo/jev-reflex-autonomy-lab","category":"High-Frequency & Simulation","plainSummary":"jev-reflex-autonomy-lab：Jev 作为 System 1 反射层，为每架无人机从 HOLD、转向、加减速等飞行动作中选择下一个最佳动作。","plainSummaryEn":"Multi-drone autonomy lab demonstrating TypeSafe Jev reflex decisions with optional System 2 strategy guidance.","jevDecisionPoint":"Jev 作为 System 1 反射层，为每架无人机从 HOLD、转向、加减速等飞行动作中选择下一个最佳动作。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"在连续交互中观察决策效果；频率依实际运行而定。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["games-simulation","typed-decisions"],"stars":12,"forks":2,"openIssues":0,"license":null,"lastCommitAt":"2026-09-20T19:31:13Z","createdAt":"2026-09-18T19:46:20Z","metadataFetchedAt":"2026-09-21T22:11:00.057Z","avatarUrl":"https://avatars.githubusercontent.com/u/32692718?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/khordoo/jev-reflex-autonomy-lab/blob/22fd32637383ac11b69e186c327c6c203cf5999d/lib/reflex/jev-server.ts","note":"自动发现的 Jev 集成证据"}]},{"id":"panachy:jevvy","name":"jevvy","author":"PanAchy","url":"https://github.com/PanAchy/jevvy","category":"Security & Guardrails","plainSummary":"为 OpenCode 等编码Agent自动放行无害的 shell 命令，对不确定命令保留人工审核。","plainSummaryEn":"Jev-powered plugins for coding agents","jevDecisionPoint":"Jev 用四个 Noul 问题判断 shell 命令是否有害、敏感、不可信或被遮蔽，以决定自动放行还是继续人工确认。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"把语义判断接到已有的命令行工作流。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["security-guardrails","typed-decisions"],"stars":13,"forks":0,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-21T02:18:02Z","createdAt":"2026-09-18T21:41:30Z","metadataFetchedAt":"2026-09-21T22:11:03.055Z","avatarUrl":"https://avatars.githubusercontent.com/u/19510452?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/PanAchy/jevvy/blob/f682078c036213e855ff6cc2fadbb119d7da7550/packages/core/src/typesafe.ts","note":"自动发现的 Jev 集成证据"}]},{"id":"lykycy123:robojev","name":"RoboJEV","author":"lykycy123","url":"https://github.com/lykycy123/RoboJEV","category":"High-Frequency & Simulation","plainSummary":"**在 MuJoCo 中，通过两阶段 JEV 控制 Franka Panda 完成三种操作任务。**","plainSummaryEn":"RoboJEV is a small, inspectable robotics laboratory. JEV receives **structured simulator state, not images**, selects an immediate intent, then selects X/Y/Z directions and a gripper command. A Cartesian controller executes the action using real MuJoCo contacts. Each task has independent physical success checks; model answers cannot declare success.","jevDecisionPoint":"Jev 根据实测仿真状态先选择任务意图，再为末端执行器选择 X/Y/Z 方向和夹爪开合指令。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"在连续交互中观察决策效果；频率依实际运行而定。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["games-simulation","typed-decisions"],"stars":24,"forks":1,"openIssues":0,"license":"Apache-2.0","lastCommitAt":"2026-09-21T11:18:53Z","createdAt":"2026-09-20T09:36:44Z","metadataFetchedAt":"2026-09-21T22:11:06.123Z","avatarUrl":"https://avatars.githubusercontent.com/u/94739529?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/lykycy123/RoboJEV/blob/aa82be09603032b0e196f2e4aaa7277d54eb7b42/src/jev_vla_sim/policy.py","note":"自动发现的 Jev 集成证据"}]},{"id":"ray-hughes:jevalyn","name":"jevalyn","author":"Ray-Hughes","url":"https://github.com/Ray-Hughes/jevalyn","category":"SDK & Decision Frameworks","plainSummary":"Jevalyn 是 Rails 应用的决策层，封装 Jev System One API，用 noul、choice 和 score 三种类型化问题返回结构化判定，并提供 Guardrail 和 Router 辅助控制流程。","plainSummaryEn":"The decision layer for your Rails app. A Rails-native wrapper around TypeSafe's Jev System One API: typed, calibrated decisions in your control flow.","jevDecisionPoint":"Jev 根据输入 state 对 noul、choice、score 类型问题给出类型化答案，用于决定紧急度、归属部门和严重程度等业务路由。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"给现有流程增加可检查的判断环节。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["multilanguage-sdk","typed-decisions"],"stars":16,"forks":1,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-21T02:08:25Z","createdAt":"2026-09-20T02:05:38Z","metadataFetchedAt":"2026-09-21T22:11:09.059Z","avatarUrl":"https://avatars.githubusercontent.com/u/5935246?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/Ray-Hughes/jevalyn/blob/98d877dd88c5664e536f358d572267671348562c/spec/jevalyn/client_spec.rb","note":"自动发现的 Jev 集成证据"}]},{"id":"win4r:jev-security-scan","name":"jev-security-scan","author":"win4r","url":"https://github.com/win4r/jev-security-scan","category":"Security & Guardrails","plainSummary":"输出文件和行号、脱敏后的证据、风险类别、模型概率及未扫描范围。支持 Codex、Claude Code，也可以作为独立 Python 命令行工具使用。","plainSummaryEn":"Reviews Agent Skills and MCP configurations and source code with local static checks and TypeSafe Jev before installation or execution, reporting file and line evidence, risk categories, model probabilities, and coverage gaps.","jevDecisionPoint":"Jev 对每个代码分块判断是否属于九类安全风险并复核证据位置与执行上下文以决定风险结论。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"把语义判断接到已有的命令行工作流。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["security-guardrails","typed-decisions"],"stars":10,"forks":1,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-19T10:09:19Z","createdAt":"2026-09-19T10:09:22Z","metadataFetchedAt":"2026-09-21T22:11:12.045Z","avatarUrl":"https://avatars.githubusercontent.com/u/42172631?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/win4r/jev-security-scan/blob/a4ea18d8e2cb5cc8447f063c9870cebe46d7663f/scripts/jev_client.py","note":"自动发现的 Jev 集成证据"}]},{"id":"kiln-ai:jev_jsonschema","name":"jev_jsonschema","author":"Kiln-AI","url":"https://github.com/Kiln-AI/jev_jsonschema","category":"SDK & Decision Frameworks","plainSummary":"该库将 JSON Schema 转换为 Jev 问题集，调用 SystemOne API 后再将答案解码为符合原 Schema 的 JSON。","plainSummaryEn":"`probabilities` is keyed by your schema's values, not Jev's internal labels, so a score of `1`–`5` reads as `\"1\"`–`\"5\"` and not `\"0\"`–`\"4\"`. Noul questions carry no confidence of their own, so `confidence` is `None` for booleans and numbers.","jevDecisionPoint":"Jev 针对输入 state 回答由 Schema 属性转换而来的每个问题（枚举选择、布尔判断、整数打分），决定各字段的取值与概率分布。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"把语义判断接到已有的命令行工作流。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["multilanguage-sdk","typed-decisions"],"stars":7,"forks":0,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-19T19:16:09Z","createdAt":"2026-09-19T17:50:30Z","metadataFetchedAt":"2026-09-21T22:11:15.183Z","avatarUrl":"https://avatars.githubusercontent.com/u/178670964?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/Kiln-AI/jev_jsonschema/blob/fccea8c2da0edf1bdfd1de30f8664c6684ba6c63/src/jev_jsonschema/client.py","note":"自动发现的 Jev 集成证据"}]},{"id":"metalbear-co:jev-auto-approve","name":"jev-auto-approve","author":"metalbear-co","url":"https://github.com/metalbear-co/jev-auto-approve","category":"Security & Guardrails","plainSummary":"这是一个 GitHub Action，调用 Jev 并行询问多个是否问题，仅当所有置信度达标时自动批准 PR，否则跳过并评论分数。","plainSummaryEn":"Jev is a decision model: it answers a typed question with a calibrated probability rather than\nprose. This action asks it one yes/no question per thing worth being sure about — answered in\nparallel in a single call — and approves only when every one of them clears your threshold:","jevDecisionPoint":"Jev 对每个问题给出 yes/no 校准概率，判断 PR 是否可合并、测试是否充分、是否需要人工审查，只有全部达到阈值才批准。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"增加一组可记录、可对照的判断信号。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["security-guardrails","cli-git-gates"],"stars":7,"forks":1,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-20T13:16:37Z","createdAt":"2026-09-20T07:47:57Z","metadataFetchedAt":"2026-09-21T22:11:18.122Z","avatarUrl":"https://avatars.githubusercontent.com/u/99899594?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/metalbear-co/jev-auto-approve/blob/5c58e7b3a3c9977a7e646504190457773d1d87fa/src/jev.mjs","note":"自动发现的 Jev 集成证据"}]},{"id":"peu77:jevfind","name":"JevFind","author":"Peu77","url":"https://github.com/Peu77/JevFind","category":"Data & Search","plainSummary":"这是一个由 Jev 驱动的语义代码搜索工具，可用自然语言查询找到相关文件、行范围和代码片段。","plainSummaryEn":"Fast semantic code search powered by Jev. Find the relevant files, line ranges, and snippets","jevDecisionPoint":"Jev 负责判断每个文件路径和代码窗口与用户自然语言概念相关的概率。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"把语义判断接到已有的命令行工作流。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["search-retrieval","cli-git-gates"],"stars":4,"forks":0,"openIssues":1,"license":"MIT","lastCommitAt":"2026-09-20T13:26:47Z","createdAt":"2026-09-19T15:48:52Z","metadataFetchedAt":"2026-09-21T22:11:21.141Z","avatarUrl":"https://avatars.githubusercontent.com/u/60301119?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/Peu77/JevFind/blob/c286f60a61d24b0f4249b2ea88495a4f52dfd51e/src/main.rs","note":"自动发现的 Jev 集成证据"}]},{"id":"shuhansun:jev-oas-sentinel","name":"jev-oas-sentinel","author":"ShuhanSun","url":"https://github.com/ShuhanSun/jev-oas-sentinel","category":"CLI & Pipelines","plainSummary":"该工具比较两个 OpenAPI 文档，用确定性检查发现结构兼容性问题，并用 TypeSafe Jev 评估变更描述中的语义风险。","plainSummaryEn":"JEV never writes a review or changes a specification. It returns typed decisions and probabilities; deterministic Python code decides whether to pass, request review, or block.","jevDecisionPoint":"Jev 对新旧 OpenAPI 契约片段的兼容性种类、受影响语义维度、旧承诺是否保留及迁移负担进行分类与打分。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"增加一组可记录、可对照的判断信号。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["cli-git-gates","typed-decisions"],"stars":3,"forks":2,"openIssues":0,"license":"Apache-2.0","lastCommitAt":"2026-09-20T19:22:17Z","createdAt":"2026-09-19T19:18:24Z","metadataFetchedAt":"2026-09-21T22:11:24.153Z","avatarUrl":"https://avatars.githubusercontent.com/u/12295127?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/ShuhanSun/jev-oas-sentinel/blob/807bf63033e5910f12fbb52b2c84dbcb127dde8f/src/jev_oas_sentinel/jev.py","note":"自动发现的 Jev 集成证据"}]},{"id":"scienthoon:jev-ood-calibration","name":"jev-ood-calibration","author":"scienthoon","url":"https://github.com/scienthoon/jev-ood-calibration","category":"SDK & Decision Frameworks","plainSummary":"该仓库用规则生成的客服工单和三个公开基准来检验 Jev 返回概率的校准情况，并提供数据生成、评测与复现脚本及原始结果。","plainSummaryEn":"Independent calibration test of TypeSafe's Jev on a task it cannot have seen: 900 rule-generated support tickets (choice / score / boolean) plus 3 public benchmarks via Vercel AI Gateway. Raw responses, ECE with noise floor, temperature refit, per-type sign of miscalibration. Reproducible for ~$0.06.","jevDecisionPoint":"Jev对每张合成支持工单的所属队列、是否愤怒和优先级分别做出choice、boolean、score决策并返回校准概率。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"在现有程序中复用接入代码，减少重复处理接口细节。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["multilanguage-sdk","typed-decisions"],"stars":3,"forks":0,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-19T07:44:28Z","createdAt":"2026-09-19T07:33:22Z","metadataFetchedAt":"2026-09-21T22:11:27.056Z","avatarUrl":"https://avatars.githubusercontent.com/u/46613837?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/scienthoon/jev-ood-calibration/blob/e092d3f1ef8baa5fd213443862b04d1db6fb63aa/scripts/jev_eval.mjs","note":"自动发现的 Jev 集成证据"}]},{"id":"kolawong:fast-compaction-dsh","name":"fast-compaction-dsh","author":"kolawong","url":"https://github.com/kolawong/fast-compaction-dsh","category":"Context GC & Filter","plainSummary":"DSH 的上下文压缩插件","plainSummaryEn":"Verdict-based context compaction for DeepSeek Harness — replaces lossy LLM summaries with fast keep/truncate/drop decisions from jev-latest; everything kept stays verbatim. Port of tamaratran/fast-jev-compaction.","jevDecisionPoint":"用 jev-latest 对每个工具调用做快速的 保留/截断/删除 判定，替换掉有损的压缩摘要——保留的内容逐字不动，任何东西都不会被改写。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"增加一组可记录、可对照的判断信号。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["context-compaction","typed-decisions"],"stars":2,"forks":0,"openIssues":0,"license":null,"lastCommitAt":"2026-09-20T17:32:37Z","createdAt":"2026-09-20T14:37:51Z","metadataFetchedAt":"2026-09-21T22:11:30.106Z","avatarUrl":"https://avatars.githubusercontent.com/u/33373300?v=4","summarySource":"readme-extractive","claimStatus":"优先保留投稿者与仓库原文，缺失语言自动补充；自动检查仅确认 Jev 集成证据，未经本站运行或性能复测。","claimStatusEn":"Author and repository text is preserved where clear; missing languages are enriched automatically. Checks establish source-level Jev integration, not runtime, safety, or performance validation.","evidence":[{"url":"https://github.com/kolawong/fast-compaction-dsh/blob/322dc84518c3ea541c22182f954e42b93f1bec88/README.md","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/kolawong/fast-compaction-dsh/blob/322dc84518c3ea541c22182f954e42b93f1bec88/README.zh.md","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/kolawong/fast-compaction-dsh/blob/322dc84518c3ea541c22182f954e42b93f1bec88/src/jev.ts","note":"固定版本的 Jev 集成与说明来源"}]},{"id":"kyu1204:jgrep","name":"jgrep","author":"kyu1204","url":"https://github.com/kyu1204/jgrep","category":"CLI & Pipelines","plainSummary":"jgrep 是用英文描述代码行为来搜索代码、git diff 和表格行的命令行工具，它把代码块与每个块一个 Noul 问题一起发给 Jev，并输出命中的 file:line，同时支持用 Noul/Choice/Score 问题检查 diff 和为 CSV/JSONL 的每一行打分。","plainSummaryEn":"grep for what code does, not what it's called. Semantic code search powered by TypeSafe Jev.","jevDecisionPoint":"Jev 对每个代码块/差异块或表格行逐一判断是否符合英文描述（Noul二值判断，Rows模式还支持Choice多选和Score打分）。","jevDecisionPointEn":"Measured (2026-09-19, jev-1.13.0): a 896-chunk TypeScript `src/` tree in 1.8 s for $0.010 (240k input tokens); repeat query 0 s from cache.","highlightBenefit":"把语义判断接进已有的数据查询流程。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["cli-git-gates","search-retrieval","coding-agents"],"stars":15,"forks":3,"openIssues":1,"license":"MIT","lastCommitAt":"2026-09-21T14:21:39Z","createdAt":"2026-09-19T04:24:17Z","metadataFetchedAt":"2026-09-21T22:11:33.056Z","avatarUrl":"https://avatars.githubusercontent.com/u/33680799?v=4","summarySource":"readme-extractive","claimStatus":"优先保留投稿者与仓库原文，缺失语言自动补充；自动检查仅确认 Jev 集成证据，未经本站运行或性能复测。","claimStatusEn":"Author and repository text is preserved where clear; missing languages are enriched automatically. Checks establish source-level Jev integration, not runtime, safety, or performance validation.","evidence":[{"url":"https://github.com/kyu1204/jgrep/blob/e91569b6c871717ac47ac1e94e4387110ac900f1/README.md","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/kyu1204/jgrep/blob/e91569b6c871717ac47ac1e94e4387110ac900f1/src/cli.ts","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/kyu1204/jgrep/blob/e91569b6c871717ac47ac1e94e4387110ac900f1/src/init.ts","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/kyu1204/jgrep/blob/e91569b6c871717ac47ac1e94e4387110ac900f1/src/jgrep.test.ts","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/kyu1204/jgrep/blob/e91569b6c871717ac47ac1e94e4387110ac900f1/src/jgrep.ts","note":"固定版本的 Jev 集成与说明来源"}]},{"id":"keltokhy:jlink","name":"jlink","author":"keltokhy","url":"https://github.com/keltokhy/jlink","category":"Data & Search","plainSummary":"jlink 让用户用 plain English 写出匹配规则，在本地做候选 blocking 生成记录对，再用 Jev Noul 逐对判断是否为同一实体并给出概率，然后解析链接并提供审计与引用文本。","plainSummaryEn":"The string baselines are best-match Jaro-Winkler and best-match TF-IDF cosine; the table shows\nthe better of the two. Exact matching after normalization scores 0.26, 0.41, 0.00, 0.00 and\n0.22.","jevDecisionPoint":"Jev 根据用户用自然语言给出的匹配规则，判断两个候选记录是否指向同一实体并返回匹配概率。","jevDecisionPointEn":"Jev judges whether two candidate records refer to the same entity under a user-supplied match definition; local code generates candidates and resolves accepted matches.","highlightBenefit":"增加一组可记录、可对照的判断信号。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["database-vector","domain-workflows","typed-decisions"],"stars":4,"forks":0,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-21T03:16:43Z","createdAt":"2026-09-18T11:11:29Z","metadataFetchedAt":"2026-09-21T22:11:36.074Z","avatarUrl":"https://avatars.githubusercontent.com/u/83777601?v=4","summarySource":"readme-extractive","claimStatus":"优先保留投稿者与仓库原文，缺失语言自动补充；自动检查仅确认 Jev 集成证据，未经本站运行或性能复测。","claimStatusEn":"Author and repository text is preserved where clear; missing languages are enriched automatically. Checks establish source-level Jev integration, not runtime, safety, or performance validation.","evidence":[{"url":"https://github.com/keltokhy/jlink/blob/65533df0b531283f2980374ada18c0fd74b6c7ab/README.md","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/keltokhy/jlink/blob/65533df0b531283f2980374ada18c0fd74b6c7ab/src/jlink/judge.py","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/keltokhy/jlink/blob/65533df0b531283f2980374ada18c0fd74b6c7ab/src/jlink/__init__.py","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/keltokhy/jlink/blob/65533df0b531283f2980374ada18c0fd74b6c7ab/src/jlink/__main__.py","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/keltokhy/jlink/blob/65533df0b531283f2980374ada18c0fd74b6c7ab/src/jlink/assets/review.js","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/keltokhy/jlink/blob/65533df0b531283f2980374ada18c0fd74b6c7ab/src/jlink/audit.py","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/keltokhy/jlink/blob/65533df0b531283f2980374ada18c0fd74b6c7ab/src/jlink/block.py","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/keltokhy/jlink/blob/65533df0b531283f2980374ada18c0fd74b6c7ab/src/jlink/cli.py","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/keltokhy/jlink/blob/65533df0b531283f2980374ada18c0fd74b6c7ab/src/jlink/core.py","note":"固定版本的 Jev 集成与说明来源"}]},{"id":"keltokhy:jsort","name":"jsort","author":"keltokhy","url":"https://github.com/keltokhy/jsort","category":"CLI & Pipelines","plainSummary":"按自然语言描述的维度对文本行/段落/文件进行两两比较排序并输出分数与标准误的命令行工具。","plainSummaryEn":"Ranks text along a plain-English criterion using pairwise Jev Noul comparisons and a locally fitted Bradley-Terry scale.","jevDecisionPoint":"Jev 对给定的自然语言维度判断文本 A 是否比文本 B 排名更高。","jevDecisionPointEn":"Jev judges whether text A ranks higher than text B on the supplied criterion; local code schedules comparisons and fits a Bradley-Terry scale with standard errors.","highlightBenefit":"增加一组可记录、可对照的判断信号。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["cli-git-gates","classification-ranking","typed-decisions"],"stars":9,"forks":0,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-21T02:40:07Z","createdAt":"2026-09-19T18:35:25Z","metadataFetchedAt":"2026-09-21T22:11:39.108Z","avatarUrl":"https://avatars.githubusercontent.com/u/83777601?v=4","summarySource":"readme-extractive","claimStatus":"优先保留投稿者与仓库原文，缺失语言自动补充；自动检查仅确认 Jev 集成证据，未经本站运行或性能复测。","claimStatusEn":"Author and repository text is preserved where clear; missing languages are enriched automatically. Checks establish source-level Jev integration, not runtime, safety, or performance validation.","evidence":[{"url":"https://github.com/keltokhy/jsort/blob/380dfe57760241681c4743bece352ea4c5086c39/README.md","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/keltokhy/jsort/blob/380dfe57760241681c4743bece352ea4c5086c39/src/jsort/model.py","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/keltokhy/jsort/blob/380dfe57760241681c4743bece352ea4c5086c39/src/jsort/__init__.py","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/keltokhy/jsort/blob/380dfe57760241681c4743bece352ea4c5086c39/src/jsort/__main__.py","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/keltokhy/jsort/blob/380dfe57760241681c4743bece352ea4c5086c39/src/jsort/cli.py","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/keltokhy/jsort/blob/380dfe57760241681c4743bece352ea4c5086c39/src/jsort/core.py","note":"固定版本的 Jev 集成与说明来源"}]},{"id":"kerpopule:hermes-jev-skills","name":"hermes-jev-skills","author":"kerpopule","url":"https://github.com/kerpopule/hermes-jev-skills","category":"Routing & Cost Optimization","plainSummary":"该仓库为 Hermes Agent 提供基于 Jev 的八个 SKILL.md 技能，负责模型路由、记忆筛选、转录压缩、技能选择以及计算机和浏览器操作决策。","plainSummaryEn":"Jev-powered model routing, memory, compaction, skill selection, computer and browser use for Hermes agents (also Claude Code and Codex)","jevDecisionPoint":"Jev 决定每一轮用哪个模型、加载哪个技能、保留哪些记忆片段与对话轮次，以及下一步 GUI/浏览器操作。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"按任务分配模型资源；具体成本收益需看项目测试。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["llm-routing-cost","typed-decisions"],"stars":387,"forks":35,"openIssues":6,"license":"MIT","lastCommitAt":"2026-09-21T17:01:49Z","createdAt":"2026-09-18T20:01:29Z","metadataFetchedAt":"2026-09-21T22:11:42.060Z","avatarUrl":"https://avatars.githubusercontent.com/u/26120791?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/kerpopule/hermes-jev-skills/blob/29e64b57026eb040b355c260a2480015bd459611/jevkit/client.py","note":"自动发现的 Jev 集成证据"}]},{"id":"win4r:jev-skill-suggester","name":"jev-skill-suggester","author":"win4r","url":"https://github.com/win4r/jev-skill-suggester","category":"CLI & Pipelines","plainSummary":"用 TypeSafe Jev 为当前任务推荐一个合适的已安装 Skill。先阅读技能描述筛选，再核对候选正文片段；允许返回“无需技能”或“不确定”。用户明确指定的技能通过本地查找优先处理。","plainSummaryEn":"A Python CLI and Codex Skill that recommends a suitable installed skill for the current task using TypeSafe Jev Choice and Noul checks without executing candidate skills.","jevDecisionPoint":"Jev根据任务文本对已安装技能的描述和正文片段做Choice排序，并用Noul判断是否值得推荐及推荐哪一个。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"把语义判断接到已有的命令行工作流。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["cli-git-gates","typed-decisions"],"stars":30,"forks":3,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-19T15:31:32Z","createdAt":"2026-09-19T15:31:35Z","metadataFetchedAt":"2026-09-21T22:11:45.085Z","avatarUrl":"https://avatars.githubusercontent.com/u/42172631?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/win4r/jev-skill-suggester/blob/05fbd7ce9ec74a2f193c09276a8d4d077d9d5e5c/scripts/jev_client.py","note":"自动发现的 Jev 集成证据"}]},{"id":"stas4000:jev-linkmap","name":"jev-linkmap","author":"stas4000","url":"https://github.com/stas4000/jev-linkmap","category":"Domain & Vertical Tools","plainSummary":"该项目抓取网站内容并生成候选内部链接，由 Jev 判断是否值得链接及选择原文中的锚文本，并支持用深模型复核分歧来改写 rubric。","plainSummaryEn":"Site: www.bles-software.com, 566 pages, 8,460 link decisions (15 candidate targets per page). Run on 19 Sep 2026. Every number below is from the run files in `out/` and `runs/`.","jevDecisionPoint":"Jev对每个候选目标判断是否值得加内链，并从正文已有短语中选择最诚实的锚文本或选无。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"把语义判断接到已有的命令行工作流。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["domain-workflows","typed-decisions"],"stars":19,"forks":0,"openIssues":0,"license":null,"lastCommitAt":"2026-09-19T19:24:00Z","createdAt":"2026-09-19T17:57:35Z","metadataFetchedAt":"2026-09-21T22:11:48.104Z","avatarUrl":"https://avatars.githubusercontent.com/u/16101773?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/stas4000/jev-linkmap/blob/58f636ecc4fb3339026e8a4f0cb49e5273f756a3/jev_linkmap/jev.py","note":"自动发现的 Jev 集成证据"}]},{"id":"jaibhasin:jev-yt-time-saver","name":"jev-yt-time-saver","author":"jaibhasin","url":"https://github.com/jaibhasin/jev-yt-time-saver","category":"Browser & OS Action","plainSummary":"jev-yt-time-saver：Jev 根据视频标题、频道、时长、描述和搜索意图判断每个 YouTube 视频是否浪费时间并给出浪费分数。","plainSummaryEn":"This project integrates Jev to provide structured decisions for its workflow. See the repository for implementation details.","jevDecisionPoint":"Jev 根据视频标题、频道、时长、描述和搜索意图判断每个 YouTube 视频是否浪费时间并给出浪费分数。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"把选择和打分接进现有程序；暂无可核验的性能对照。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["browser-automation","typed-decisions"],"stars":5,"forks":0,"openIssues":0,"license":null,"lastCommitAt":"2026-09-20T12:31:52Z","createdAt":"2026-09-19T17:35:36Z","metadataFetchedAt":"2026-09-21T22:11:51.170Z","avatarUrl":"https://avatars.githubusercontent.com/u/66554426?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/jaibhasin/jev-yt-time-saver/blob/6ceeb82396bd5187a3dea61ab159c32fdb7552be/background/background.js","note":"自动发现的 Jev 集成证据"}]},{"id":"reindeer-ai:pi-jev-guard","name":"pi-jev-guard","author":"Reindeer-AI","url":"https://github.com/Reindeer-AI/pi-jev-guard","category":"Security & Guardrails","plainSummary":"该 Pi 扩展在写入前使用 TypeSafe Jev 检查提议的代码修改是否违反 Markdown 规则，并返回违规规则原文与行号范围。","plainSummaryEn":"Check Pi code edits against repository Markdown rules with TypeSafe Jev","jevDecisionPoint":"Jev 判断提议的代码修改是否引入或加重了对每条 Markdown 规则块的违反。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"把下一步调查集中到更相关的证据上。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["security-guardrails","typed-decisions"],"stars":5,"forks":0,"openIssues":0,"license":null,"lastCommitAt":"2026-09-19T14:07:55Z","createdAt":"2026-09-19T12:23:03Z","metadataFetchedAt":"2026-09-21T22:11:54.127Z","avatarUrl":"https://avatars.githubusercontent.com/u/218930453?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/Reindeer-AI/pi-jev-guard/blob/144ed4cb5d0066b2149c91d4035367d609214647/src/jev.ts","note":"自动发现的 Jev 集成证据"}]},{"id":"larguesa:jev-search","name":"jev-search","author":"larguesa","url":"https://github.com/larguesa/jev-search","category":"Data & Search","plainSummary":"这是一个用 Jev 按语义逐行检索文本的 Python 命令行工具，用作关键词搜索的补充视角。","plainSummaryEn":"Experimental semantic line search with TypeSafe Jev via OpenRouter. Python CLI with no runtime dependencies.","jevDecisionPoint":"Jev 对每一行 l{i} 判定其是否满足查询语义意图，返回 noul 概率分数。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"把语义判断接进已有的数据查询流程。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["search-retrieval","cli-git-gates","typed-decisions"],"stars":6,"forks":1,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-20T15:14:53Z","createdAt":"2026-09-19T15:08:07Z","metadataFetchedAt":"2026-09-21T22:11:57.076Z","avatarUrl":"https://avatars.githubusercontent.com/u/25781395?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/larguesa/jev-search/blob/8aa403554d7904d3c14224cfda02a4e2144c0ccd/jev_search.py","note":"自动发现的 Jev 集成证据"}]},{"id":"imohitmayank:jevfill","name":"jevfill","author":"imohitmayank","url":"https://github.com/imohitmayank/jevfill","category":"Browser & OS Action","plainSummary":"Jevfill 是一个 Chrome 扩展，可调用 Jev 根据用户粘贴的纯文本笔记自动填写网页表单。","plainSummaryEn":"Open `test/sample-form.html` in the browser, configure the extension, and click **Autofill page**.","jevDecisionPoint":"Jev 为每个扫描到的表单字段从用户笔记行中选择最匹配的一行并返回置信度。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"把语义判断接到已有的命令行工作流。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["browser-automation","typed-decisions"],"stars":15,"forks":0,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-20T18:15:04Z","createdAt":"2026-09-20T14:33:23Z","metadataFetchedAt":"2026-09-21T22:12:00.110Z","avatarUrl":"https://avatars.githubusercontent.com/u/1883779?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/imohitmayank/jevfill/blob/12246f784a49f5e48be9aca8c2dd7588f1d1fa9f/src/jev/client.ts","note":"自动发现的 Jev 集成证据"}]},{"id":"thanh-mathieu95:jev-model-tokengate","name":"jev-model-tokengate","author":"Thanh-Mathieu95","url":"https://github.com/Thanh-Mathieu95/jev-model-tokengate","category":"Security & Guardrails","plainSummary":"该项目是一个 OpenAI 兼容的流式代理，在Token到达用户前用滑动缓冲加并行评估进行拦截，实现零泄漏的内容过滤。","plainSummaryEn":"An OpenAI-compatible proxy that sits between your LLM and your users. It evaluates each\nsliding window of tokens **while the response is still streaming** and cuts the stream\n**before** a violating token can reach the screen.","jevDecisionPoint":"Jev 对每个待释放的滑动窗口文本并行给出5个安全标准的0-1 noul分数，以决定放行该批次还是丢弃缓冲并切断流。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"给现有流程增加可检查的判断环节。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["security-guardrails","typed-decisions"],"stars":4,"forks":0,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-20T13:09:40Z","createdAt":"2026-09-20T03:07:18Z","metadataFetchedAt":"2026-09-21T22:12:03.118Z","avatarUrl":"https://avatars.githubusercontent.com/u/98025755?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/Thanh-Mathieu95/jev-model-tokengate/blob/3072b7362dd7fbd8de53fb254975ee43bbd4b0a2/evaluator.js","note":"自动发现的 Jev 集成证据"}]},{"id":"virolea:jev","name":"jev","author":"virolea","url":"https://github.com/virolea/jev","category":"SDK & Decision Frameworks","plainSummary":"该 Gem 是 Typesafe Jev 模型 API 的 Ruby 客户端，可在一次查询中并行提出多个问题并读取 noul、choice 和 score 类型的答案。","plainSummaryEn":"Ruby client for the typesafe AI Jev model","jevDecisionPoint":"Jev 根据传入的 state 并行判断 ask 的真值概率、choose 的最优选项和 score 的等级分布。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"把语义判断接到已有的命令行工作流。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["multilanguage-sdk","typed-decisions"],"stars":3,"forks":0,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-20T19:58:53Z","createdAt":"2026-09-19T14:01:41Z","metadataFetchedAt":"2026-09-21T22:12:06.155Z","avatarUrl":"https://avatars.githubusercontent.com/u/3525369?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/virolea/jev/blob/d15edc9646541ac2a20182397c9a4df5319c883e/lib/jev/client.rb","note":"自动发现的 Jev 集成证据"}]},{"id":"alexj11324:open-jev-approvals","name":"open-jev-approvals","author":"alexj11324","url":"https://github.com/alexj11324/open-jev-approvals","category":"Security & Guardrails","plainSummary":"agent 要执行工具时，hook 会拦住这次调用，交给 TypeSafe Jev 审查，再由\n本地政策给出 `allow` 或 `deny`。走第三方 API 时，harness 自带的审批往往\n不可用，这个门就是那种场景下的 auto mode。Jev 负责审查，本地政策负责\n裁决。Jev 给不出结论时放行——拒绝必须有「这个动作确实危险」的正面证据。","plainSummaryEn":"Binary approval gate for Codex and Claude Code — every intercepted tool call is reviewed by TypeSafe JEV and composed through a versioned local policy, with scoped authorization.","jevDecisionPoint":"每次拦截到工具调用时，由Jev回答风险、授权、危害和范围等固定问题以决定放行还是拒绝。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"给现有流程增加可检查的判断环节。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["security-guardrails","typed-decisions"],"stars":3,"forks":0,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-20T09:52:39Z","createdAt":"2026-09-20T05:17:00Z","metadataFetchedAt":"2026-09-21T22:12:09.136Z","avatarUrl":"https://avatars.githubusercontent.com/u/183887963?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/alexj11324/open-jev-approvals/blob/889d06f3e2b6225132788a028fe4d3f35f0ee2ae/internal/jev/client.go","note":"自动发现的 Jev 集成证据"}]},{"id":"cobusgreyling:jev","name":"Jev","author":"cobusgreyling","url":"https://github.com/cobusgreyling/Jev","category":"SDK & Decision Frameworks","plainSummary":"该仓库是 Jev 的非官方展示与操作实验台，用智能家居演示、交互实验、示例脚本、TypeScript harness CLI 和 Agent skills 说明 Choice、Score、Noul 并行判断与置信度路由的用法。","plainSummaryEn":"Unofficial TypeSafe Jev showcase — System One decisions, not chat.","jevDecisionPoint":"Jev 对同一 state 并行执行 Choice/Score/Noul 类型化判定并返回置信度，由代码据此做路由与执行。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"按任务分配模型资源；具体成本收益需看项目测试。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["multilanguage-sdk","typed-decisions","security-guardrails"],"stars":3,"forks":0,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-20T07:32:02Z","createdAt":"2026-09-19T10:35:48Z","metadataFetchedAt":"2026-09-21T22:12:12.075Z","avatarUrl":"https://avatars.githubusercontent.com/u/7868717?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/cobusgreyling/Jev/blob/c636087edf4100a097a456e559c4b2fa769b80e8/jev_lab/client.py","note":"自动发现的 Jev 集成证据"}]},{"id":"zhangxaochen:dsh-jev","name":"dsh-jev","author":"zhangxaochen","url":"https://github.com/zhangxaochen/dsh-jev","category":"Security & Guardrails","plainSummary":"它补上 dsh 自己没有的那一层语义判断：引入 ~150ms 极低延迟的非生成式决策原语（Noul、Choice、Score），做**动态工具剪枝**（省 Prompt Token、降首字延迟）、**语义死循环阻断**与**高危执行安全门禁**。判定走 System One 而非生成式采样，所以快、可复现，且成本可忽略：**每次判定 ≈ $0.00013**（2026-09-20 实测 2,254 次判决，每次输入 12.4KiB，按 $0.042/M 输入 token 计费、输出免费）。","plainSummaryEn":"Jev (System One decision model) plugin suite for DeepSeek Harness (dsh)","jevDecisionPoint":"Jev 负责判断工具相关性排序、轨迹是否陷入语义死循环以及高风险 shell/文件操作是否放行。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"给现有流程增加可检查的判断环节。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["security-guardrails","typed-decisions"],"stars":3,"forks":0,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-21T07:06:59Z","createdAt":"2026-09-18T10:36:35Z","metadataFetchedAt":"2026-09-21T22:12:15.054Z","avatarUrl":"https://avatars.githubusercontent.com/u/3416677?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/zhangxaochen/dsh-jev/blob/8204830bfa5132283db1a732a275f31a6da8bc29/lib/typesafe-client.js","note":"自动发现的 Jev 集成证据"}]},{"id":"ishantanu:jevmetrics","name":"jevmetrics","author":"ishantanu","url":"https://github.com/ishantanu/jevmetrics","category":"CLI & Pipelines","plainSummary":"jevmetrics 是一个 OpenTelemetry Collector 指标处理器，它调用 Jev 模型根据指标元数据推断操作价值，并据此注释或过滤指标。","plainSummaryEn":"Use it to assess unfamiliar instrumentation, review candidates for reduced retention, and selectively filter metrics before they reach a primary backend. Inference runs asynchronously, and cached assessments let subsequent batches use the same decision without another API call.","jevDecisionPoint":"Jev根据指标元数据推断指标的操作价值、冗余度、保留概率和处置建议。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"增加一组可记录、可对照的判断信号。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["cli-git-gates","typed-decisions"],"stars":3,"forks":0,"openIssues":2,"license":"Apache-2.0","lastCommitAt":"2026-09-21T21:50:32Z","createdAt":"2026-09-19T12:42:07Z","metadataFetchedAt":"2026-09-21T22:12:18.104Z","avatarUrl":"https://avatars.githubusercontent.com/u/6175822?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/ishantanu/jevmetrics/blob/111dcba8582e9a5156ef7ae515c1ec40f6bc7409/internal/evaluator/jev.go","note":"自动发现的 Jev 集成证据"}]},{"id":"gudcks0305:jev-java","name":"jev-java","author":"gudcks0305","url":"https://github.com/gudcks0305/jev-java","category":"SDK & Decision Frameworks","plainSummary":"该项目是 TypeSafe Jev、OpenRouter 和 Vercel AI Gateway 的非官方 Java SDK，可将应用状态作为 Choice、Noul 和 Score 类型化问题批量提交并返回类型化 Jev 判断结果，同时提供 Spring Boot 自动配置和 WebClient 传输支持。","plainSummaryEn":"Unofficial Java SDK for TypeSafe Jev and Vercel AI Gateway, with Spring Boot and WebClient support","jevDecisionPoint":"根据输入状态文本由Jev做出部门路由选择、紧急与否的概率判断和严重程度评分。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"在现有程序中复用接入代码，减少重复处理接口细节。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["multilanguage-sdk","typed-decisions"],"stars":3,"forks":0,"openIssues":6,"license":"MIT","lastCommitAt":"2026-09-21T14:32:43Z","createdAt":"2026-09-19T15:26:15Z","metadataFetchedAt":"2026-09-21T22:12:21.162Z","avatarUrl":"https://avatars.githubusercontent.com/u/84506708?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/gudcks0305/jev-java/blob/2d2b5c26c803802717b344e5dd5fc4e5524ba4e0/jev-spring-boot-autoconfigure/src/main/java/io/github/gudcks0305/jev/autoconfigure/JevAutoConfiguration.java","note":"自动发现的 Jev 集成证据"}]},{"id":"joshhu:jevtest","name":"jevtest","author":"joshhu","url":"https://github.com/joshhu/jevtest","category":"Domain & Vertical Tools","plainSummary":"情緒測謊器：嘴上說「好」，心裡真的好嗎？用 TypeSafe Jev（System One 模型）透過 OpenRouter 即時判斷，並與一般 LLM 對照","plainSummaryEn":"It provides a web and CLI demo that sends user text to Jev through OpenRouter for Choice, Score, and Noul probability judgments and compares the results side-by-side with a general LLM.","jevDecisionPoint":"Jev 判断对方回复背后的真实含义、火气等级以及是否需要马上主动关心。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"把语义判断接到已有的命令行工作流。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["domain-workflows","typed-decisions"],"stars":4,"forks":0,"openIssues":0,"license":null,"lastCommitAt":"2026-09-20T14:25:13Z","createdAt":"2026-09-20T11:02:36Z","metadataFetchedAt":"2026-09-21T22:12:24.107Z","avatarUrl":"https://avatars.githubusercontent.com/u/10594453?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/joshhu/jevtest/blob/7b20148292223df96b4345d66bb2691ff511a79d/main.py","note":"自动发现的 Jev 集成证据"}]},{"id":"omkarghugarkar007:actiongate-jev","name":"actiongate-jev","author":"omkarghugarkar007","url":"https://github.com/omkarghugarkar007/actiongate-jev","category":"Security & Guardrails","plainSummary":"ActionGate 是面向 AI Agent 工具调用的授权网关，结合确定性策略与 Jev 评估提议动作，并对获批的精确动作签发一次性许可，在 MCP 与 HTTP 执行边界进行核销。","plainSummaryEn":"Open-source Jev tool-calling authorization gateway for AI agents: deterministic policy, exact-action single-use permits, MCP and HTTP enforcement.","jevDecisionPoint":"Jev 判断Agent提议的工具调用是否与用户意图一致，为是否签发一次性许可提供证据。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"给现有流程增加可检查的判断环节。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["security-guardrails","typed-decisions"],"stars":2,"forks":0,"openIssues":3,"license":"Apache-2.0","lastCommitAt":"2026-09-20T18:42:56Z","createdAt":"2026-09-19T06:20:33Z","metadataFetchedAt":"2026-09-21T22:12:27.056Z","avatarUrl":"https://avatars.githubusercontent.com/u/62425457?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/omkarghugarkar007/actiongate-jev/blob/cd358fc0c2490c70586e0d7b73c2788af0c9599b/packages/decision-provider/src/typesafe-jev.ts","note":"自动发现的 Jev 集成证据"}]},{"id":"jkrup:jeveryword","name":"jeveryword","author":"jkrup","url":"https://github.com/jkrup/jeveryword","category":"Data & Search","plainSummary":"该库将文本切分为带编号的词供 Jev 选择，并将选中的编号还原为原文子串，用于字段抽取、词标注和原文引用。","plainSummaryEn":"Jev answers multiple-choice questions and does not generate text, so on its own it cannot\nreturn a name, an email address or a quote. jeveryword numbers the words of your text, offers\nthose numbers as the answer options, and converts the numbers Jev picks back into the original\nsubstring with its character offsets.","jevDecisionPoint":"Jev 负责从编号词选项中选择每个字段的起止词、定位所属句子并为每个词选择标签。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"把选择和打分接进现有程序；暂无可核验的性能对照。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["search-retrieval","classification-ranking"],"stars":2,"forks":0,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-20T18:31:48Z","createdAt":"2026-09-19T23:26:55Z","metadataFetchedAt":"2026-09-21T22:12:30.072Z","avatarUrl":"https://avatars.githubusercontent.com/u/519731?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/jkrup/jeveryword/blob/6fed1518bbb84cc4865401630f37c01dc3decc10/src/client.mjs","note":"自动发现的 Jev 集成证据"}]},{"id":"adkid-zephyr:work-with-jev","name":"work-with-jev","author":"Adkid-Zephyr","url":"https://github.com/Adkid-Zephyr/work-with-jev","category":"Domain & Vertical Tools","plainSummary":"**工作群消息太多？用 Jev 挑出需要你处理的事，整理成一份跨群待办。**","plainSummaryEn":"Work with Jev is a local-first message classifier that uses Jev to sort work messages into urgent, to-do, worth-reading, and skippable groups with cross-chat to-do management and Feishu and WeCom adapters.","jevDecisionPoint":"Jev 判断每条工作消息是否与当前用户相关、是否需行动、是否紧急、有无参考价值，并返回概率以分入四类。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"把语义判断接到已有的命令行工作流。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["domain-workflows","typed-decisions"],"stars":3,"forks":1,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-20T17:53:45Z","createdAt":"2026-09-20T13:12:43Z","metadataFetchedAt":"2026-09-21T22:12:33.088Z","avatarUrl":"https://avatars.githubusercontent.com/u/169631528?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/Adkid-Zephyr/work-with-jev/blob/8d2db18ef753cd84e06a75dcaa6fa881b35c5a77/server.mjs","note":"自动发现的 Jev 集成证据"}]},{"id":"minilv:jev-auto-router","name":"Jev-Auto-Router","author":"miniLV","url":"https://github.com/miniLV/Jev-Auto-Router","category":"Routing & Cost Optimization","plainSummary":"架构中，Jev 负责每次调用的模型与推理档位选择；本地 Responses 代理负责保持 Codex 会话和工具循环连续；任务结束后独立验收。Router Compass 把选路、实际用量和验收结果放在一起，回答一个问题：**少用旗舰模型之后，任务是否仍然正确完成，整体开销是否真的下降？**","plainSummaryEn":"Jev Auto Router (Jev Router): experimental per-call GPT model routing for Codex via TypeSafe Jev and a local Responses proxy, with independent task verification.","jevDecisionPoint":"Jev 为每次有意义的模型调用从宿主可请求的（模型, 推理档位）候选对中做一次 Choice 选择。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"按任务分配模型资源；具体成本收益需看项目测试。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["llm-routing-cost","typed-decisions"],"stars":3,"forks":0,"openIssues":0,"license":"Apache-2.0","lastCommitAt":"2026-09-20T15:47:37Z","createdAt":"2026-08-01T02:04:30Z","metadataFetchedAt":"2026-09-21T22:12:36.069Z","avatarUrl":"https://avatars.githubusercontent.com/u/20677007?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/miniLV/Jev-Auto-Router/blob/9a75f14fda63b554b05abb9b696d62eec3b795ff/src/index.ts","note":"自动发现的 Jev 集成证据"}]},{"id":"completetech-llc-ai-research:jev-311-heatmap","name":"jev-311-heatmap","author":"CompleteTech-LLC-AI-Research","url":"https://github.com/CompleteTech-LLC-AI-Research/jev-311-heatmap","category":"Data & Search","plainSummary":"该项目下载 NYC 311 投诉数据，用 JEV 评估投诉描述，并在地理网格上生成可交互热力图。","plainSummaryEn":"The live run excluded 205 reports with missing or invalid coordinates, completed **634 API calls without retries**, and reported **539,979 input tokens**. Repeated descriptions share one evaluation.","jevDecisionPoint":"JEV根据311投诉描述判断潜在影响严重程度、安全隐患概率和所属问题类别。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"增加一组可记录、可对照的判断信号。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["search-retrieval","typed-decisions"],"stars":2,"forks":0,"openIssues":0,"license":null,"lastCommitAt":"2026-09-20T00:24:57Z","createdAt":"2026-09-20T00:20:04Z","metadataFetchedAt":"2026-09-21T22:12:39.166Z","avatarUrl":"https://avatars.githubusercontent.com/u/201413662?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/CompleteTech-LLC-AI-Research/jev-311-heatmap/blob/172eae5dfb40abc72590b5451e5330edb4c53c91/generate_heatmap.py","note":"自动发现的 Jev 集成证据"}]},{"id":"lucianfialho:jev-model-router","name":"jev-model-router","author":"lucianfialho","url":"https://github.com/lucianfialho/jev-model-router","category":"Routing & Cost Optimization","plainSummary":"该库使用 Jev 对请求进行分类，并从 OpenRouter 实时模型目录中筛选出满足约束的最便宜可用模型。","plainSummaryEn":"Cost-optimized OpenRouter model router using TypeSafe's Jev, with a live full-catalog scorer instead of a hardcoded model list","jevDecisionPoint":"Jev 负责对用户请求进行领域、复杂度和长上下文/视觉/延迟需求的分类，以决定选择哪个价位的模型。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"按任务分配模型资源；具体成本收益需看项目测试。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["llm-routing-cost","typed-decisions"],"stars":2,"forks":0,"openIssues":1,"license":"MIT","lastCommitAt":"2026-09-19T20:11:55Z","createdAt":"2026-09-19T20:08:43Z","metadataFetchedAt":"2026-09-21T22:12:42.081Z","avatarUrl":"https://avatars.githubusercontent.com/u/3296497?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/lucianfialho/jev-model-router/blob/4a16d03957df28bd5989060266da80191456301c/src/model_router/classifier.py","note":"自动发现的 Jev 集成证据"}]},{"id":"anisselbd:jev-phishing-bench","name":"jev-phishing-bench","author":"anisselbd","url":"https://github.com/anisselbd/jev-phishing-bench","category":"Security & Guardrails","plainSummary":"jev-phishing-bench：Jev 负责判断每封邮件是否为钓鱼邮件、链接是否可点击，并对五个钓鱼信号给出概率。","plainSummaryEn":"The signal result above was challenged on three points: no non-AI baseline, selection and evaluation on the same\nemails, and no equivalent decomposition for the LLM. Three controls were added (`bench/heuristics.py`,\n`bench/protocol.py`, `run_llm_signals.py`); nothing above was changed. Full tables in `results/report.md`, chart in\n`results/controls.png`.","jevDecisionPoint":"Jev 负责判断每封邮件是否为钓鱼邮件、链接是否可点击，并对五个钓鱼信号给出概率。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"增加一组可记录、可对照的判断信号。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["security-guardrails","evaluation-benchmarks","typed-decisions"],"stars":3,"forks":0,"openIssues":1,"license":null,"lastCommitAt":"2026-09-17T14:16:41Z","createdAt":"2026-09-16T22:42:44Z","metadataFetchedAt":"2026-09-21T22:12:45.050Z","avatarUrl":"https://avatars.githubusercontent.com/u/142441328?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/anisselbd/jev-phishing-bench/blob/1d56e8c64d029a9554a0874e2ef2901ed196e230/run_jev.py","note":"自动发现的 Jev 集成证据"}]},{"id":"ajanm007:jevrag","name":"jevrag","author":"ajanm007","url":"https://github.com/ajanm007/jevrag","category":"SDK & Decision Frameworks","plainSummary":"该项目为 RAG 流程提供可插拔的决策层，通过统一的 state → Decision → confidence → action 接口实现检索停止、分块、上下文选择、回答拒答和缓存信任五类决策，并以 Jev 作为首个可替换后端，同时提供校准评估工具。","plainSummaryEn":"Replaces hardcoded RAG thresholds with explicit calibrated decision points. Five primitives (retrieval stopping, chunk splitting, context selection, answer abstention, cache trust) behind one swappable state → Decision → confidence → action interface, each evaluated on real datasets with a calibration harness that reports honestly.","jevDecisionPoint":"Jev 根据检索状态对证据充分性、块边界、上下文选择、答案弃答和缓存信任五个 RAG 决策点给出类型化答案与置信度。","jevDecisionPointEn":"jevrag/decision.py — the `ask(state, questions) -> DecisionResult` method calls `api.typesafe.ai/v1/systemone` via the Jev backend. All five primitives in `jevrag/primitives/` route through this method.","highlightBenefit":"减少后续处理的冗余信息。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["multilanguage-sdk","typed-decisions","evaluation-benchmarks"],"stars":1,"forks":0,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-21T16:38:22Z","createdAt":"2026-09-20T09:52:52Z","metadataFetchedAt":"2026-09-21T22:12:48.102Z","avatarUrl":"https://avatars.githubusercontent.com/u/120266872?v=4","summarySource":"readme-extractive","claimStatus":"优先保留投稿者与仓库原文，缺失语言自动补充；自动检查仅确认 Jev 集成证据，未经本站运行或性能复测。","claimStatusEn":"Author and repository text is preserved where clear; missing languages are enriched automatically. Checks establish source-level Jev integration, not runtime, safety, or performance validation.","evidence":[{"url":"https://github.com/ajanm007/jevrag/blob/2b5563a9a4a68c592e727fc390890798dd5be2fb/README.md","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/ajanm007/jevrag/blob/2b5563a9a4a68c592e727fc390890798dd5be2fb/jevrag/adapters/rag_jev_selector.py","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/ajanm007/jevrag/blob/2b5563a9a4a68c592e727fc390890798dd5be2fb/scripts/eval_rag_jev.py","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/ajanm007/jevrag/blob/2b5563a9a4a68c592e727fc390890798dd5be2fb/scripts/eval_scifact_ragjev.py","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/ajanm007/jevrag/blob/2b5563a9a4a68c592e727fc390890798dd5be2fb/jevrag/decision.py","note":"固定版本的 Jev 集成与说明来源"}]},{"id":"hemanth:jevish","name":"jevish","author":"hemanth","url":"https://github.com/hemanth/jevish","category":"SDK & Decision Frameworks","plainSummary":"jevish 是一个 JavaScript 语义模式匹配与零样本判断库，使用 Jev 的 Choice 做标签分类、使用 Noul 做布尔判定，并支持本地与云端 Jev 协同执行。","plainSummaryEn":"Every mode auto-curries when called with only the patterns:","jevDecisionPoint":"当本地零依赖引擎置信度不足时，由云端 Jev System One 裁决零样本标签多选一和布尔谓词真假。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"把语义判断接到已有的命令行工作流。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["multilanguage-sdk","classification-ranking","typed-decisions"],"stars":1,"forks":0,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-21T13:19:01Z","createdAt":"2026-09-19T08:38:51Z","metadataFetchedAt":"2026-09-21T22:12:51.092Z","avatarUrl":"https://avatars.githubusercontent.com/u/18315?v=4","summarySource":"readme-extractive","claimStatus":"优先保留投稿者与仓库原文，缺失语言自动补充；自动检查仅确认 Jev 集成证据，未经本站运行或性能复测。","claimStatusEn":"Author and repository text is preserved where clear; missing languages are enriched automatically. Checks establish source-level Jev integration, not runtime, safety, or performance validation.","evidence":[{"url":"https://github.com/hemanth/jevish/blob/8766b550901b07cf080dc6c0e61ccf325d35f047/README.md","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/hemanth/jevish/blob/8766b550901b07cf080dc6c0e61ccf325d35f047/lib/engine.js","note":"固定版本的 Jev 集成与说明来源"}]},{"id":"hemanth:traffic-guard","name":"traffic-guard","author":"hemanth","url":"https://github.com/hemanth/traffic-guard","category":"Security & Guardrails","plainSummary":"该项目是为传入 HTTP 请求提供流量分类与拦截判断的网关，提供 Node.js 与 Python 实现，并可选用 TypeSafe System One 通过 Noul、Choice 和 Score 进行语义评估。","plainSummaryEn":"High-throughput traffic and attack defense gate for incoming HTTP traffic with zero required dependencies, wire-order header validation, and TypeSafe System One acceleration for bot mitigation, exploit detection, and risk scoring.","jevDecisionPoint":"Jev 并行回答 is_bot/is_attack/is_spoofed 是否拦截、traffic_type 分类多选一、risk_level 风险打分，以决定放行/挑战/限速/拦截。","jevDecisionPointEn":"traffic-guard progressively evaluates requests against TypeSafe System One using Noul (`is_bot`, `is_attack`, `is_spoofed`), Choice (`traffic_type`), and Score (`risk_level`).","highlightBenefit":"把语义判断接到已有的命令行工作流。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["security-guardrails","typed-decisions"],"stars":1,"forks":0,"openIssues":0,"license":null,"lastCommitAt":"2026-09-21T12:50:56Z","createdAt":"2026-09-18T06:51:34Z","metadataFetchedAt":"2026-09-21T22:12:54.122Z","avatarUrl":"https://avatars.githubusercontent.com/u/18315?v=4","summarySource":"readme-extractive","claimStatus":"优先保留投稿者与仓库原文，缺失语言自动补充；自动检查仅确认 Jev 集成证据，未经本站运行或性能复测。","claimStatusEn":"Author and repository text is preserved where clear; missing languages are enriched automatically. Checks establish source-level Jev integration, not runtime, safety, or performance validation.","evidence":[{"url":"https://github.com/hemanth/traffic-guard/blob/192673bd801866827a81ecfb19413f0e74a90990/README.md","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/hemanth/traffic-guard/blob/192673bd801866827a81ecfb19413f0e74a90990/node/src/gate.ts","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/hemanth/traffic-guard/blob/192673bd801866827a81ecfb19413f0e74a90990/node/src/policy.ts","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/hemanth/traffic-guard/blob/192673bd801866827a81ecfb19413f0e74a90990/python/traffic_guard/gate.py","note":"固定版本的 Jev 集成与说明来源"}]},{"id":"hemanth:jev-chess","name":"jev-chess","author":"hemanth","url":"https://github.com/hemanth/jev-chess","category":"High-Frequency & Simulation","plainSummary":"该项目使用 TypeSafe AI System One 的 Choice、Score 和 Noul 原语，将自然语言走棋意图解析为合法走法，并提供局面评估、历史棋手风格对手和整局对局分类。","plainSummaryEn":"Chess moves, evaluations, persona opponents, and game classification using TypeSafe AI System One models. Resolves natural language move intents into legal moves, evaluates positional sharpness and king risk in parallel, and powers historical persona opponents (Tal, Capablanca, Petrosian).","jevDecisionPoint":"Jev 负责将自然语言走棋意图解析为合法走法，并对走法的锐度、战略主题、王翼攻击风险和棋风人格维度进行打分与选择。","jevDecisionPointEn":"jev-chess invokes Jev primitives via `@typesafe-ai/sdk` for natural-language move resolution, multi-dimensional position evaluations, and persona style scoring.","highlightBenefit":"把语义判断接到已有的命令行工作流。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["games-simulation","typed-decisions","evaluation-benchmarks"],"stars":2,"forks":0,"openIssues":1,"license":null,"lastCommitAt":"2026-09-21T12:51:28Z","createdAt":"2026-09-19T04:29:34Z","metadataFetchedAt":"2026-09-21T22:12:57.104Z","avatarUrl":"https://avatars.githubusercontent.com/u/18315?v=4","summarySource":"readme-extractive","claimStatus":"优先保留投稿者与仓库原文，缺失语言自动补充；自动检查仅确认 Jev 集成证据，未经本站运行或性能复测。","claimStatusEn":"Author and repository text is preserved where clear; missing languages are enriched automatically. Checks establish source-level Jev integration, not runtime, safety, or performance validation.","evidence":[{"url":"https://github.com/hemanth/jev-chess/blob/2e35170801a0374119f1862209f3cd818c00f467/README.md","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/hemanth/jev-chess/blob/2e35170801a0374119f1862209f3cd818c00f467/src/typeSafeClient.ts","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/hemanth/jev-chess/blob/2e35170801a0374119f1862209f3cd818c00f467/public/app.js","note":"固定版本的 Jev 集成与说明来源"}]},{"id":"hemanth:tc39-atlas","name":"tc39-atlas","author":"hemanth","url":"https://github.com/hemanth/tc39-atlas","category":"Domain & Vertical Tools","plainSummary":"该项目是交互式 TC39 提案语义浏览工具，使用 Jev 按采用路径、认知负担、Web 兼容风险和意图原型对 ECMAScript 提案进行分类展示。","plainSummaryEn":"Interactive semantic explorer and taxonomy for TC39 proposals. Applies TypeSafe AI System One (Jev) to classify ECMAScript proposals across adoption pathways, cognitive overhead, web-compatibility risk, and foundational intent archetypes.","jevDecisionPoint":"Jev 对每个 TC39 提案在采用路径、复杂度、动机、Web 兼容风险和意图原型等维度上执行 Choice 多选、Score 打分和 Noul 门禁判断。","jevDecisionPointEn":"tc39-atlas enriches ECMAScript proposals using TypeSafe System One (Jev) across multi-dimensional rubrics (domain, complexity score, adoption choice, cognitive overhead score, web-compat risk score, and intent archetypes).","highlightBenefit":"把语义判断接到已有的命令行工作流。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["domain-workflows","classification-ranking","typed-decisions"],"stars":1,"forks":0,"openIssues":0,"license":null,"lastCommitAt":"2026-09-17T06:51:58Z","createdAt":"2026-09-17T06:27:38Z","metadataFetchedAt":"2026-09-21T22:13:00.137Z","avatarUrl":"https://avatars.githubusercontent.com/u/18315?v=4","summarySource":"readme-extractive","claimStatus":"优先保留投稿者与仓库原文，缺失语言自动补充；自动检查仅确认 Jev 集成证据，未经本站运行或性能复测。","claimStatusEn":"Author and repository text is preserved where clear; missing languages are enriched automatically. Checks establish source-level Jev integration, not runtime, safety, or performance validation.","evidence":[{"url":"https://github.com/hemanth/tc39-atlas/blob/fef00f6f3707a0b52b21dc08b769f7be5ccf08f4/README.md","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/hemanth/tc39-atlas/blob/fef00f6f3707a0b52b21dc08b769f7be5ccf08f4/src/lib/typesafeEngine.ts","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/hemanth/tc39-atlas/blob/fef00f6f3707a0b52b21dc08b769f7be5ccf08f4/src/server/typesafeApi.ts","note":"固定版本的 Jev 集成与说明来源"}]},{"id":"hemanth:pkg-gate","name":"pkg-gate","author":"hemanth","url":"https://github.com/hemanth/pkg-gate","category":"Security & Guardrails","plainSummary":"pkg-gate 是在安装前评估 npm 生命周期脚本的门禁工具，它使用 TypeSafe System One 以 Choice 判断意图、以 Score 评估威胁严重程度、以 Noul 判断机密访问和远程执行，并输出 allow、warn 或 block 的判定结果。","plainSummaryEn":"Pre-install security gate for npm lifecycle scripts using TypeSafe System One. Evaluates preinstall, install, and postinstall hooks across intent, threat severity, secret access, and remote execution to intercept supply-chain attacks before execution.","jevDecisionPoint":"对每个 preinstall/install/postinstall 脚本并行判定意图分类、威胁评分与窃密/远程执行概率，以决定放行、告警或拦截。","jevDecisionPointEn":"pkg-gate uses `@typesafe-ai/sdk` to evaluate lifecycle scripts in parallel across intent (Choice), threat severity (Score), secret exfiltration (Noul), and remote execution (Noul).","highlightBenefit":"把语义判断接到已有的命令行工作流。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["security-guardrails","cli-git-gates","typed-decisions"],"stars":0,"forks":0,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-21T12:50:21Z","createdAt":"2026-09-17T01:47:58Z","metadataFetchedAt":"2026-09-21T22:13:03.116Z","avatarUrl":"https://avatars.githubusercontent.com/u/18315?v=4","summarySource":"readme-extractive","claimStatus":"优先保留投稿者与仓库原文，缺失语言自动补充；自动检查仅确认 Jev 集成证据，未经本站运行或性能复测。","claimStatusEn":"Author and repository text is preserved where clear; missing languages are enriched automatically. Checks establish source-level Jev integration, not runtime, safety, or performance validation.","evidence":[{"url":"https://github.com/hemanth/pkg-gate/blob/7649d07768b3682c1d9d9ef94f0b6232c1e2f8d8/README.md","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/hemanth/pkg-gate/blob/7649d07768b3682c1d9d9ef94f0b6232c1e2f8d8/src/gate.js","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/hemanth/pkg-gate/blob/7649d07768b3682c1d9d9ef94f0b6232c1e2f8d8/src/index.d.ts","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/hemanth/pkg-gate/blob/7649d07768b3682c1d9d9ef94f0b6232c1e2f8d8/src/index.js","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/hemanth/pkg-gate/blob/7649d07768b3682c1d9d9ef94f0b6232c1e2f8d8/src/questions.js","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/hemanth/pkg-gate/blob/7649d07768b3682c1d9d9ef94f0b6232c1e2f8d8/src/resolver.js","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/hemanth/pkg-gate/blob/7649d07768b3682c1d9d9ef94f0b6232c1e2f8d8/src/simulator.js","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/hemanth/pkg-gate/blob/7649d07768b3682c1d9d9ef94f0b6232c1e2f8d8/src/tui.js","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/hemanth/pkg-gate/blob/7649d07768b3682c1d9d9ef94f0b6232c1e2f8d8/src/utils.js","note":"固定版本的 Jev 集成与说明来源"}]},{"id":"hemanth:hfjev","name":"hfjev","author":"hemanth","url":"https://github.com/hemanth/hfjev","category":"Data & Search","plainSummary":"hfjev 可加载 Hugging Face 数据集，并使用 Choice、Score 和 Noul 等 Jev 原语对每一行进行类型化语义分类。","plainSummaryEn":"Classify Hugging Face datasets across typed semantic dimensions with TypeSafe Jev System One. Auto-adapts evaluation rubrics to dataset domains (reviews, news, LLM tuning, support) and classifies rows in a single parallel System One call with calibrated probabilities.","jevDecisionPoint":"Jev对每条Hugging Face数据集文本在情感、主题、深度、剧透等类型化维度上并行做出Choice多选、Score打分和Noul门禁判定。","jevDecisionPointEn":"hfjev classifies dataset rows in parallel across typed Choice, Score, and Noul dimensions against TypeSafe System One.","highlightBenefit":"把语义判断接到已有的命令行工作流。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["search-retrieval","classification-ranking","evaluation-benchmarks"],"stars":1,"forks":0,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-21T17:00:26Z","createdAt":"2026-09-20T10:08:36Z","metadataFetchedAt":"2026-09-21T22:13:06.106Z","avatarUrl":"https://avatars.githubusercontent.com/u/18315?v=4","summarySource":"readme-extractive","claimStatus":"优先保留投稿者与仓库原文，缺失语言自动补充；自动检查仅确认 Jev 集成证据，未经本站运行或性能复测。","claimStatusEn":"Author and repository text is preserved where clear; missing languages are enriched automatically. Checks establish source-level Jev integration, not runtime, safety, or performance validation.","evidence":[{"url":"https://github.com/hemanth/hfjev/blob/122245bedc52af5e73e21cb26263e4d174313ce7/README.md","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/hemanth/hfjev/blob/122245bedc52af5e73e21cb26263e4d174313ce7/js/bin/hfjev.js","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/hemanth/hfjev/blob/122245bedc52af5e73e21cb26263e4d174313ce7/playground/src/components/ApiKeyModal.tsx","note":"固定版本的 Jev 集成与说明来源"}]},{"id":"hemanth:tool-prune","name":"tool-prune","author":"hemanth","url":"https://github.com/hemanth/tool-prune","category":"Routing & Cost Optimization","plainSummary":"该项目为 AI Agent 提供校准化的工具选择与 schema 裁剪功能，可在调用 LLM 前筛选出相关候选工具，并支持离线 TurboQuant 与云端 TypeSafe System One（Jev）双引擎。","plainSummaryEn":"Calibrated tool selection and schema pruning for AI agents. Dual-engine: zero-dependency offline TurboQuant or TypeSafe System One (Jev). Prunes candidate MCP tools and schemas down to the relevant set before calling LLMs to eliminate hallucinations and save tokens.","jevDecisionPoint":"Jev 对用户意图做原子决策：用 choice 选出最匹配的工具并给出校准概率，用 noul 判断是否需要开放式创造性生成。","jevDecisionPointEn":"Jev makes atomic decisions on tool selection (`choice`) and whether user intent requires open-ended creative generation (`noul`), with calibrated probability distributions.","highlightBenefit":"让现有 Agent 通过通用接口使用 Jev。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["llm-routing-cost","mcp-integrations","typed-decisions"],"stars":5,"forks":0,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-21T12:49:51Z","createdAt":"2026-09-17T14:27:29Z","metadataFetchedAt":"2026-09-21T22:13:09.104Z","avatarUrl":"https://avatars.githubusercontent.com/u/18315?v=4","summarySource":"readme-extractive","claimStatus":"优先保留投稿者与仓库原文，缺失语言自动补充；自动检查仅确认 Jev 集成证据，未经本站运行或性能复测。","claimStatusEn":"Author and repository text is preserved where clear; missing languages are enriched automatically. Checks establish source-level Jev integration, not runtime, safety, or performance validation.","evidence":[{"url":"https://github.com/hemanth/tool-prune/blob/e55c586901a4cb33cb138a856cee74869430af9e/README.md","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/hemanth/tool-prune/blob/e55c586901a4cb33cb138a856cee74869430af9e/js/lib/router.js","note":"固定版本的 Jev 集成与说明来源"}]},{"id":"zjunlp:jevloop","name":"JevLoop","author":"zjunlp","url":"https://github.com/zjunlp/JevLoop","category":"Routing & Cost Optimization","plainSummary":"**你的 agent loop 里每一个岔路口都是一次完整的大模型调用。而它们没有一个是「生成」。**","plainSummaryEn":"The agent loop where decisions don't cost a large language model call. Zero deps, runs offline, no API key needed.","jevDecisionPoint":"Jev 负责裁决 Agent 循环中的每个分叉：是否需行动、选哪个工具、风险定级与是否需人工授权、步骤是否成功、是否完成及能否交付。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"增加一组可记录、可对照的判断信号。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["llm-routing-cost","typed-decisions"],"stars":11,"forks":3,"openIssues":1,"license":"Apache-2.0","lastCommitAt":"2026-09-21T15:07:13Z","createdAt":"2026-09-20T15:21:09Z","metadataFetchedAt":"2026-09-21T22:13:12.094Z","avatarUrl":"https://avatars.githubusercontent.com/u/41887875?v=4","summarySource":"readme-extractive","claimStatus":"优先保留投稿者与仓库原文，缺失语言自动补充；自动检查仅确认 Jev 集成证据，未经本站运行或性能复测。","claimStatusEn":"Author and repository text is preserved where clear; missing languages are enriched automatically. Checks establish source-level Jev integration, not runtime, safety, or performance validation.","evidence":[{"url":"https://github.com/zjunlp/JevLoop/blob/c6024d9a304c79205c1996e94504de1075834144/README.md","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/zjunlp/JevLoop/blob/c6024d9a304c79205c1996e94504de1075834144/README.zh-CN.md","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/zjunlp/JevLoop/blob/c6024d9a304c79205c1996e94504de1075834144/src/agent.ts","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/zjunlp/JevLoop/blob/c6024d9a304c79205c1996e94504de1075834144/src/backends.ts","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/zjunlp/JevLoop/blob/c6024d9a304c79205c1996e94504de1075834144/src/decision-compile.ts","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/zjunlp/JevLoop/blob/c6024d9a304c79205c1996e94504de1075834144/src/decision-shape.ts","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/zjunlp/JevLoop/blob/c6024d9a304c79205c1996e94504de1075834144/src/decisiondoc.ts","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/zjunlp/JevLoop/blob/c6024d9a304c79205c1996e94504de1075834144/src/decisions.ts","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/zjunlp/JevLoop/blob/c6024d9a304c79205c1996e94504de1075834144/src/policy.ts","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/zjunlp/JevLoop/blob/c6024d9a304c79205c1996e94504de1075834144/src/budget.ts","note":"固定版本的 Jev 集成与说明来源"}]},{"id":"4rays:profanity-checker","name":"profanity-checker","author":"4rays","url":"https://github.com/4rays/profanity-checker","category":"Security & Guardrails","plainSummary":"profanity-checker：Jev 判断输入文本是否含亵渎语言，以及用户名是否含字面或伪装（谐音/形近）的亵渎内容。","plainSummaryEn":"Deploy this Worker once, then call it from your other Workers.","jevDecisionPoint":"Jev 判断输入文本是否含亵渎语言，以及用户名是否含字面或伪装（谐音/形近）的亵渎内容。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"减少后续处理的冗余信息。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["security-guardrails","typed-decisions"],"stars":1,"forks":0,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-20T21:02:58Z","createdAt":"2026-09-20T21:07:29Z","metadataFetchedAt":"2026-09-21T22:13:15.104Z","avatarUrl":"https://avatars.githubusercontent.com/u/110794587?v=4","summarySource":"readme-extractive","claimStatus":"优先保留投稿者与仓库原文，缺失语言自动补充；自动检查仅确认 Jev 集成证据，未经本站运行或性能复测。","claimStatusEn":"Author and repository text is preserved where clear; missing languages are enriched automatically. Checks establish source-level Jev integration, not runtime, safety, or performance validation.","evidence":[{"url":"https://github.com/4rays/profanity-checker/blob/0a0466f614a9c35e4697a0e8a20600e0104e8151/README.md","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/4rays/profanity-checker/blob/0a0466f614a9c35e4697a0e8a20600e0104e8151/src/endpoints/profanityCheck.ts","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/4rays/profanity-checker/blob/0a0466f614a9c35e4697a0e8a20600e0104e8151/src/index.ts","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/4rays/profanity-checker/blob/0a0466f614a9c35e4697a0e8a20600e0104e8151/src/types.ts","note":"固定版本的 Jev 集成与说明来源"}]},{"id":"achimala:jev-paint","name":"jev-paint","author":"achimala","url":"https://github.com/achimala/jev-paint","category":"Creative Tools","plainSummary":"把 Jev 的结构化判断接进程序；具体用途与决策流程请查看项目源码。","plainSummaryEn":"Requires Python 3.9+ and a modern browser with module workers and OffscreenCanvas\n(current Chrome, Edge, Firefox, or Safari). No packages, build step, or Node installation needed.","jevDecisionPoint":"Jev 为每个网格像素点的颜色类别输出概率分布，决定画面的色彩、明暗与轮廓形态。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"保留可编辑的组成部分与决策过程。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["creative-multimedia","typed-decisions"],"stars":49,"forks":4,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-19T11:57:55Z","createdAt":"2026-09-18T19:40:40Z","metadataFetchedAt":"2026-09-21T22:13:18.073Z","avatarUrl":"https://avatars.githubusercontent.com/u/1448366?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/achimala/jev-paint/blob/ecf9c48d290e086bde790d6e24b93324667155e9/server.py","note":"自动发现的 Jev 集成证据"}]},{"id":"ethanplusai:jev-chat-for-twitch","name":"jev-chat-for-twitch","author":"ethanplusai","url":"https://github.com/ethanplusai/jev-chat-for-twitch","category":"Context GC & Filter","plainSummary":"这是一个 Chrome 扩展，可在 Twitch 直播旁增加一列只显示经 Jev 筛选的消息。","plainSummaryEn":"Filter any live Twitch chat with Jev: a bring-your-own-key Chrome extension","jevDecisionPoint":"Jev 对每条 Twitch 聊天消息进行分类打分，判断其是否属于用户所选意图以决定是否展示。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"把选择和打分接进现有程序；暂无可核验的性能对照。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["context-compaction","browser-automation","typed-decisions"],"stars":10,"forks":0,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-19T15:50:49Z","createdAt":"2026-09-19T15:30:38Z","metadataFetchedAt":"2026-09-21T22:13:21.087Z","avatarUrl":"https://avatars.githubusercontent.com/u/249879463?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/ethanplusai/jev-chat-for-twitch/blob/80d33c4b0f39d471e817414caff06b6d1984515f/extension/jev.mjs","note":"自动发现的 Jev 集成证据"}]},{"id":"promptengineer48:laya-vs-jev-arena","name":"laya-vs-jev-arena","author":"PromptEngineer48","url":"https://github.com/PromptEngineer48/laya-vs-jev-arena","category":"High-Frequency & Simulation","plainSummary":"该项目让本地开源 Laya 与 API 驱动的 Jev 在贪吃蛇竞速和格斗对战中同场竞技，每一步动作都由模型实时决策。","plainSummaryEn":"Laya (open source, local) vs TypeSafe Jev (API): two AI models race in Snake and fight in a Mortal-Kombat-style arena. Every move is a real model decision.","jevDecisionPoint":"在贪吃蛇中决定转向方向与是否冲刺，在格斗中决定移动攻击格挡动作与是否重击。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"把语义判断接到已有的命令行工作流。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["games-simulation","typed-decisions"],"stars":13,"forks":4,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-21T06:11:48Z","createdAt":"2026-09-21T06:11:56Z","metadataFetchedAt":"2026-09-21T22:13:24.078Z","avatarUrl":"https://avatars.githubusercontent.com/u/130393461?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/PromptEngineer48/laya-vs-jev-arena/blob/858f76b9f6d55ac13b1c66bf578e72556ffd09a7/server.py","note":"自动发现的 Jev 集成证据"}]},{"id":"oldmoldycake:jev_vampire_survivors","name":"jev_vampire_survivors","author":"oldmoldycake","url":"https://github.com/oldmoldycake/jev_vampire_survivors","category":"High-Frequency & Simulation","plainSummary":"该项目让 Jev 模型通过 BepInEx 插件和 Python 大脑在原生 Linux 版 Steam 游戏 Vampire Survivors 中选择角色、关卡、升级和移动方向，并通过浏览器仪表盘实时展示决策。","plainSummaryEn":"TypeSafe's Jev model plays Vampire Survivors on Steam: BepInEx plugin + Python brain + live decision dashboard. Native Linux only.","jevDecisionPoint":"Jev根据实时游戏状态决定所选角色、关卡、每次升级选项和每秒四次的行走方向。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"把语义判断接到已有的命令行工作流。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["games-simulation","typed-decisions"],"stars":5,"forks":1,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-19T05:26:59Z","createdAt":"2026-09-18T21:46:40Z","metadataFetchedAt":"2026-09-21T22:13:27.082Z","avatarUrl":"https://avatars.githubusercontent.com/u/79052622?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/oldmoldycake/jev_vampire_survivors/blob/dae83d82593a13b882bf0abd5f481b5219357f64/brain/jev_vs/jev_client.py","note":"自动发现的 Jev 集成证据"}]},{"id":"taodav:jev_deep_rl","name":"jev_deep_rl","author":"taodav","url":"https://github.com/taodav/jev_deep_rl","category":"High-Frequency & Simulation","plainSummary":"该项目将 Jev 作为 Gymnasium 和 Atari 环境中的策略，通过游戏专用适配器把观测转换为结构化状态并选择合法动作，同时记录奖励与决策而不训练模型权重。","plainSummaryEn":"This project evaluates a fixed model. It records rewards and decisions without\ntraining or updating model weights. A seeded random policy provides a local baseline.","jevDecisionPoint":"Jev 根据结构化游戏状态通过单个 Choice 问题选择合法的下一步动作。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"把语义判断接到已有的命令行工作流。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["games-simulation","typed-decisions"],"stars":5,"forks":0,"openIssues":0,"license":null,"lastCommitAt":"2026-09-20T17:31:52Z","createdAt":"2026-09-19T17:13:05Z","metadataFetchedAt":"2026-09-21T22:13:30.089Z","avatarUrl":"https://avatars.githubusercontent.com/u/16575984?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/taodav/jev_deep_rl/blob/423133b991c6426e1dee460eb126124f9f76382d/jev_rl/cli.py","note":"自动发现的 Jev 集成证据"}]},{"id":"kolibril13:jev-in-blender-experiment","name":"jev-in-blender-experiment","author":"kolibril13","url":"https://github.com/kolibril13/jev-in-blender-experiment","category":"Creative Tools","plainSummary":"这是一个 Blender 扩展，在三维视图侧栏增加 Jev 搜索页，可用自然语言查找并执行对应的 Blender 操作。","plainSummaryEn":"Blender exposes ~2,500 operators; a TypeSafe `Choice` question holds at most 255\noptions, so the search is hierarchical (two requests per search):","jevDecisionPoint":"Jev根据用户自然语言和Blender上下文对操作模块和具体bpy.ops算子进行分层排序选择。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"把选择和打分接进现有程序；暂无可核验的性能对照。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["creative-multimedia","typed-decisions"],"stars":5,"forks":0,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-19T14:54:28Z","createdAt":"2026-09-19T11:46:09Z","metadataFetchedAt":"2026-09-21T22:13:33.150Z","avatarUrl":"https://avatars.githubusercontent.com/u/44469195?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/kolibril13/jev-in-blender-experiment/blob/878ef49607fb86ddfa88d1eb51159df85a1df39d/jev_operator_search/jev_api.py","note":"自动发现的 Jev 集成证据"}]},{"id":"collapseindex:jev-builder","name":"jev-builder","author":"collapseindex","url":"https://github.com/collapseindex/jev-builder","category":"SDK & Decision Frameworks","plainSummary":"这是一个在浏览器中填写文本和问题来生成 Jev 请求、复制请求并多次运行查看答案稳定性的表单工具。","plainSummaryEn":"A browser form for building requests to TypeSafe's Jev: pick a template, fill in the blanks, copy the request. No JSON, no install, runs locally.","jevDecisionPoint":"Jev根据用户粘贴的文本对用户自定义的是非、评分或多选项问题作出判断。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"把语义判断接到已有的命令行工作流。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["multilanguage-sdk","typed-decisions"],"stars":3,"forks":0,"openIssues":0,"license":null,"lastCommitAt":"2026-09-20T04:33:50Z","createdAt":"2026-09-19T23:17:48Z","metadataFetchedAt":"2026-09-21T22:13:36.160Z","avatarUrl":"https://avatars.githubusercontent.com/u/234891853?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/collapseindex/jev-builder/blob/b4d350c683da6d5890b97c84d1b534a57143e6a5/jev-builder-core.js","note":"自动发现的 Jev 集成证据"}]},{"id":"teempai:jev-in-codex","name":"jev-in-codex","author":"teempai","url":"https://github.com/teempai/jev-in-codex","category":"Data & Search","plainSummary":"该仓库是 Codex 插件，通过 Jev 对工作区内 JSONL 文本记录进行批量标注，生成每条记录带一个标签的输出文件以供复核。","plainSummaryEn":"Jev-powered tool and skill selection, context search, and output triage for Codex via MCP","jevDecisionPoint":"Jev 对输入 JSONL 中每条文本记录做出单选分类决策，输出唯一标签与置信度。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"增加一组可记录、可对照的判断信号。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["search-retrieval","typed-decisions","coding-agents"],"stars":3,"forks":0,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-21T05:23:02Z","createdAt":"2026-09-19T08:32:15Z","metadataFetchedAt":"2026-09-21T22:13:39.114Z","avatarUrl":"https://avatars.githubusercontent.com/u/11815232?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/teempai/jev-in-codex/blob/9f8f032918ac01d4efa1b4ee26eca0a4e27ddc08/src/labelling.ts","note":"自动发现的 Jev 集成证据"}]},{"id":"ajensenwaud:hermes-jev-plugin","name":"hermes-jev-plugin","author":"ajensenwaud","url":"https://github.com/ajensenwaud/hermes-jev-plugin","category":"SDK & Decision Frameworks","plainSummary":"该插件为 Hermes Agent 提供四个基于 TypeSafe Jev 的决策工具，分别用于二元判断、选项路由、评分和批量评估。","plainSummaryEn":"TypeSafe Jev (System One) decision tools for Hermes Agent: jev_check / jev_route / jev_score / jev_evaluate","jevDecisionPoint":"Jev 根据错误文本、代码 diff 或任务描述，对是否可重试、归属子系统、严重度等类型化问题返回带概率的结构化判定。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"把选择和打分接进现有程序；暂无可核验的性能对照。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["multilanguage-sdk","typed-decisions"],"stars":3,"forks":0,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-19T07:23:08Z","createdAt":"2026-09-19T07:23:19Z","metadataFetchedAt":"2026-09-21T22:13:42.069Z","avatarUrl":"https://avatars.githubusercontent.com/u/5827908?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/ajensenwaud/hermes-jev-plugin/blob/b3d29f71770a3447ad0b3db00658f64a8edc7dd3/client.py","note":"自动发现的 Jev 集成证据"}]},{"id":"hndrr:comfyui-jev","name":"ComfyUI-Jev","author":"hndrr","url":"https://github.com/hndrr/ComfyUI-Jev","category":"Creative Tools","plainSummary":"该项目为 ComfyUI 提供使用 Jev 进行文本选择、条件判断、评分和数字提取的自定义节点，并可将结果传递给其他节点使用。","plainSummaryEn":"Custom nodes for using Jev's text interpretation and judgments in ComfyUI. Use natural-language instructions to select candidates, evaluate conditions, score text, or extract numbers, then pass the results to other nodes. Jev judgments use the TypeSafe API by default.","jevDecisionPoint":"Jev根据自然语言指令对输入文本状态执行候选选择、条件真假判断、打分和数值提取，并为请求选择本地Skill。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"把选择和打分接进现有程序；暂无可核验的性能对照。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["creative-multimedia","typed-decisions"],"stars":4,"forks":0,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-20T06:49:46Z","createdAt":"2026-09-19T06:12:32Z","metadataFetchedAt":"2026-09-21T22:13:45.062Z","avatarUrl":"https://avatars.githubusercontent.com/u/53109993?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/hndrr/ComfyUI-Jev/blob/12239dab9f03cd4c3dead4b043028ea602b8d43b/api.py","note":"自动发现的 Jev 集成证据"}]},{"id":"charlyhno-eng:jev-codex-pilot","name":"jev-codex-pilot","author":"Charlyhno-eng","url":"https://github.com/Charlyhno-eng/jev-codex-pilot","category":"Routing & Cost Optimization","plainSummary":"JEV Codex Pilot 将待办任务转为受控开发流程，由 JEV 评估任务并推荐 Codex 模型与推理深度，同时提供 Kanban 看板、实时进度、Token 用量和 AGENTS.md 上下文管理。","plainSummaryEn":"A Codex overlay incorporating JEV to make the best decisions regarding model selection and depth of reasoning. All while automating the process via an automated Kanban system.","jevDecisionPoint":"JEV对每个任务进行任务类型分类、复杂度打分和任务精确度评估，以推荐合适的Codex模型和推理深度。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"把语义判断接到已有的命令行工作流。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["llm-routing-cost","typed-decisions"],"stars":2,"forks":0,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-21T20:28:17Z","createdAt":"2026-09-19T16:03:17Z","metadataFetchedAt":"2026-09-21T22:13:48.108Z","avatarUrl":"https://avatars.githubusercontent.com/u/184155180?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/Charlyhno-eng/jev-codex-pilot/blob/fefc7f5c5a04c6acbfcd430252f4ab4a5ab7d4f7/src/core/jev-provider.ts","note":"自动发现的 Jev 集成证据"}]},{"id":"lgy1027:jevshield","name":"jevshield","author":"lgy1027","url":"https://github.com/lgy1027/jevshield","category":"Security & Guardrails","plainSummary":"该项目是基于 Jev 的 Agent 工具调用安全门，通过单次 Choice/Noul/Score 评估拦截高风险操作，并提供本地启发式兜底和 LangChain 集成。","plainSummaryEn":"Sub-100ms security gate for AI agent tool calls, powered by TypeSafe's Jev (System-1) decision model. Single-request Choice/Noul/Score evaluation, dual-factor blocking matrix, calibrated-confidence routing, fail-closed parsing, zero-config local fallback. LangChain-ready.","jevDecisionPoint":"Jev 对每次 Agent 工具调用一次性评估风险等级、是否不可逆破坏及爆炸半径，以决定放行或拦截。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"给现有流程增加可检查的判断环节。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["security-guardrails","typed-decisions"],"stars":2,"forks":0,"openIssues":0,"license":"Apache-2.0","lastCommitAt":"2026-09-20T12:34:27Z","createdAt":"2026-09-20T12:04:26Z","metadataFetchedAt":"2026-09-21T22:13:51.100Z","avatarUrl":"https://avatars.githubusercontent.com/u/38779388?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/lgy1027/jevshield/blob/b669def84f0d7b71a8165ad1888d7fb9daf1a4ac/jevshield/client.py","note":"自动发现的 Jev 集成证据"}]},{"id":"brnyxx:jev-ra","name":"jev-ra","author":"brnyxx","url":"https://github.com/brnyxx/jev-ra","category":"Browser & OS Action","plainSummary":"**面向 CLI 编码智能体的高速浏览器操作层。** Claude Code、Codex 或任何 MCP 客户端把目标交给\njev-ra。System One 决策模型 TypeSafe Jev 在一次往返中同时选出每一步的操作和目标元素。制定计划、\n提供要输入的文本、读取页面内容、在 jev-ra 上交时接手，这些都由调用方的智能体完成。循环内不会再跑\n第二个 LLM。","plainSummaryEn":"Browser use for coding agents, 3-5x faster than browser-use. MCP server + CLI; TypeSafe Jev decides every step in ~300 ms.","jevDecisionPoint":"Jev 在每一步用一次请求同时决定操作类型、目标元素、输入值、步骤有效性与目标是否达成。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"让现有 Agent 通过通用接口使用 Jev。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["browser-automation","cli-git-gates","typed-decisions"],"stars":2,"forks":1,"openIssues":2,"license":"MIT","lastCommitAt":"2026-09-21T05:04:07Z","createdAt":"2026-09-18T02:44:55Z","metadataFetchedAt":"2026-09-21T22:13:54.156Z","avatarUrl":"https://avatars.githubusercontent.com/u/74225210?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/brnyxx/jev-ra/blob/7a966cddee04fc1c6b862ec293035bb9e380dfe9/jev_ra/decide/client.py","note":"自动发现的 Jev 集成证据"}]},{"id":"rajasekharponakala:jev-mcp","name":"jev-mcp","author":"rajasekharponakala","url":"https://github.com/rajasekharponakala/jev-mcp","category":"MCP & Integrations","plainSummary":"该项目是封装 Jev 模型的 MCP 服务器，为 Agent 提供类型化的 noul 判断、多选和评分结果。","plainSummaryEn":"MCP server wrapping TypeSafe's Jev System One models — typed noul/choice/score judgments for AI agents","jevDecisionPoint":"对输入的 state 内容执行 noul 二值判断、choice 多选分类或 score 等级评分的类型化决策。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"让现有 Agent 通过通用接口使用 Jev。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["mcp-integrations","typed-decisions"],"stars":2,"forks":0,"openIssues":0,"license":"AGPL-3.0","lastCommitAt":"2026-09-21T04:11:27Z","createdAt":"2026-09-21T04:08:49Z","metadataFetchedAt":"2026-09-21T22:13:57.085Z","avatarUrl":"https://avatars.githubusercontent.com/u/11900597?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/rajasekharponakala/jev-mcp/blob/77b97a61022695427f20cadbf988781a90733c6b/server.py","note":"自动发现的 Jev 集成证据"}]},{"id":"friedjof:jev-mobile","name":"jev-mobile","author":"Friedjof","url":"https://github.com/Friedjof/jev-mobile","category":"Browser & OS Action","plainSummary":"jev-mobile 是一个自主的 Android 子 Agent，以 TypeSafe Jev 为决策器，对 USB 连接的设备执行观察、归一化、决策、变更和验证的持久任务循环。","plainSummaryEn":"`jev-mobile` accepts a high-level task, persists it in SQLite, and lets one durable worker execute **observe → normalize → decide → mutate → verify** against a USB-connected Android device. Jev receives concise semantics and technically valid actions; it never generates coordinates, MCP calls, or arbitrary code.","jevDecisionPoint":"Jev根据语义化界面状态和有效候选动作从限定选项中决定下一步移动端操作与笔记意图分类。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"把语义判断接到已有的命令行工作流。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["browser-automation","typed-decisions"],"stars":3,"forks":3,"openIssues":1,"license":"MIT","lastCommitAt":"2026-09-18T18:59:16Z","createdAt":"2026-09-17T03:45:35Z","metadataFetchedAt":"2026-09-21T22:14:00.081Z","avatarUrl":"https://avatars.githubusercontent.com/u/60787233?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/Friedjof/jev-mobile/blob/996a49158638bcb4b4ed6564616493f86f10805c/src/jev_mobile/providers/jev.py","note":"自动发现的 Jev 集成证据"}]},{"id":"bojansandhaus:jev-decisions","name":"jev-decisions","author":"bojansandhaus","url":"https://github.com/bojansandhaus/jev-decisions","category":"Security & Guardrails","plainSummary":"该插件为 Hermes 等 Agent 提供 Jev 评审工具，用于审查风险操作、核对证据支持以及记录本地决策历史。","plainSummaryEn":"Jev Decisions Plugin for Hermes (and other AI Agents): tool risk reviews, human approval recommendations, evidence checks, and a local decision journal.","jevDecisionPoint":"Jev 根据提交的计划、证据和状态，对风险、证据支持度和选项适配度做出 noul/score/choice 类型的量化决策。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"把选择和打分接进现有程序；暂无可核验的性能对照。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["security-guardrails","typed-decisions"],"stars":2,"forks":0,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-20T09:19:29Z","createdAt":"2026-09-19T08:54:17Z","metadataFetchedAt":"2026-09-21T22:14:03.106Z","avatarUrl":"https://avatars.githubusercontent.com/u/114113320?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/bojansandhaus/jev-decisions/blob/f3a5c93392b705db217e7e04d57d0c4deb5855f6/jev_client.py","note":"自动发现的 Jev 集成证据"}]},{"id":"rorshopping:jev-browser-local","name":"jev-browser-local","author":"rorshopping","url":"https://github.com/rorshopping/jev-browser-local","category":"Browser & OS Action","plainSummary":"该项目把 jev-browser 接到本地 Jev 风格决策引擎与桥接服务，在本机完成浏览器操作决策和输入文本生成，无需调用云端模型。","plainSummaryEn":"Run jev-browser on a fully local JEV-style decision engine (no cloud API). Warm-browser fork, VRAM guard, measured benchmarks, run traces.","jevDecisionPoint":"Jev 根据当前页面状态和候选元素列表，在每一步决定点击、输入或停止等浏览器操作。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"让现有 Agent 通过通用接口使用 Jev。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["browser-automation","typed-decisions"],"stars":2,"forks":0,"openIssues":0,"license":null,"lastCommitAt":"2026-09-19T22:18:12Z","createdAt":"2026-09-18T20:03:42Z","metadataFetchedAt":"2026-09-21T22:14:06.117Z","avatarUrl":"https://avatars.githubusercontent.com/u/35996656?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/rorshopping/jev-browser-local/blob/349f70af7f8669ff3dc97fb4099304249326c229/jev-browser-fork/src/provider.ts","note":"自动发现的 Jev 集成证据"}]},{"id":"shinpr:jev-reranker","name":"jev-reranker","author":"shinpr","url":"https://github.com/shinpr/jev-reranker","category":"Data & Search","plainSummary":"该工具从标准输入读取 JSON 候选文档，并使用 Jev 按查询对其进行重排、过滤或压缩后输出。","plainSummaryEn":"Rerank, filter, and compress JSON search results with TypeSafe AI's Jev.","jevDecisionPoint":"Jev 根据查询判断每个候选文档或句子单元的相关性／证据得分，以决定排序、过滤或压缩保留。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"把语义判断接进已有的数据查询流程。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["search-retrieval","classification-ranking"],"stars":2,"forks":0,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-20T22:55:04Z","createdAt":"2026-09-20T01:42:31Z","metadataFetchedAt":"2026-09-21T22:14:09.108Z","avatarUrl":"https://avatars.githubusercontent.com/u/52246947?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/shinpr/jev-reranker/blob/1557847b47daed4f722912a9c7a0139a2fb6b538/src/http.rs","note":"自动发现的 Jev 集成证据"}]},{"id":"zhuboer:ego-jev","name":"ego-jev","author":"ZHUBoer","url":"https://github.com/ZHUBoer/ego-jev","category":"Browser & OS Action","plainSummary":"该项目让 Agent 用 Ego Lite 操作浏览器，并调用 Jev 完成语义目标选择、过滤、排序、分类和文本证据判断。","plainSummaryEn":"Complete browser tasks with Ego Lite and actively call Jev for semantic target selection, filtering, ranking, classification and text evidence judgments.","jevDecisionPoint":"Jev 负责判断观察到的哪个链接、按钮或卡片符合用户意图，并完成过滤、排序、分类和文本证据校验。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"增加一组可记录、可对照的判断信号。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["browser-automation","typed-decisions"],"stars":2,"forks":0,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-19T18:58:50Z","createdAt":"2026-09-19T18:37:21Z","metadataFetchedAt":"2026-09-21T22:14:12.168Z","avatarUrl":"https://avatars.githubusercontent.com/u/40383266?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/ZHUBoer/ego-jev/blob/c369ff61b7fe3f99047d073cbc870565af9b902c/scripts/jev.mjs","note":"自动发现的 Jev 集成证据"}]},{"id":"gholtzap:jev-codex-model-and-effort-router","name":"jev-codex-model-and-effort-router","author":"gholtzap","url":"https://github.com/gholtzap/jev-codex-model-and-effort-router","category":"Routing & Cost Optimization","plainSummary":"该项目在每次 Codex 消息时由 Jev 按请求复杂度选择 model 和 effort，并提供 macOS 菜单栏应用来管理路由偏好与可选池。","plainSummaryEn":"Copy and paste this into your coding agent:","jevDecisionPoint":"Jev 根据每次 Codex 请求的复杂度选择应使用的模型和 effort 等级。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"把选择和打分接进现有程序；暂无可核验的性能对照。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["llm-routing-cost","typed-decisions"],"stars":2,"forks":0,"openIssues":1,"license":null,"lastCommitAt":"2026-09-21T18:12:47Z","createdAt":"2026-09-18T19:24:59Z","metadataFetchedAt":"2026-09-21T22:14:15.096Z","avatarUrl":"https://avatars.githubusercontent.com/u/100532011?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/gholtzap/jev-codex-model-and-effort-router/blob/50b7edbf21d314026e29d9d3617bca7e3f01816c/jev_client.py","note":"自动发现的 Jev 集成证据"}]},{"id":"ddfeyes:jev-mode","name":"jev-mode","author":"ddfeyes","url":"https://github.com/ddfeyes/jev-mode","category":"CLI & Pipelines","plainSummary":"jev-mode 是把 Agent 批量语义判断移出上下文、交由 Jev 做类型化裁决的 Python 工具。","plainSummaryEn":"I kept watching coding agents burn context on decisions that aren't hard - triage 400 tickets, tag 600 files, route to one of six teams. jev-mode moves those verdicts to a typed-judgment model. I A/B'd it: 78% fewer tokens, 16x less work-attributable input, accuracy 96.1% vs 93.7%. Python, no deps, MIT.","jevDecisionPoint":"Jev 对每条工单/文件/Agent步骤做出类型化裁决，如归属团队、是否过期、步骤是否放行。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"减少后续处理的冗余信息。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["cli-git-gates","typed-decisions"],"stars":2,"forks":0,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-18T19:42:28Z","createdAt":"2026-09-18T14:51:43Z","metadataFetchedAt":"2026-09-21T22:14:18.097Z","avatarUrl":"https://avatars.githubusercontent.com/u/107373711?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/ddfeyes/jev-mode/blob/97615aa5cc14586488c11f6644a5b845827234c6/src/jev_mode/client.py","note":"自动发现的 Jev 集成证据"}]},{"id":"kindintelligence:jev-rust-review","name":"jev-rust-review","author":"kindintelligence","url":"https://github.com/kindintelligence/jev-rust-review","category":"MCP & Integrations","plainSummary":"这是一个为 Claude Code 提供的 Rust 代码审查插件，用 MCP 服务器运行 rustc、Clippy 与 cargo-semver-checks 并结合 Jev 对变更代码做语义审查。","plainSummaryEn":"Rust-aware code review for Claude Code and coding agents, powered by TypeSafe Jev","jevDecisionPoint":"Jev 对每个变更单元回答类型化问题，完成分流标记和缺陷验证（是否成立、严重程度、是否值得报告）。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"把下一步调查集中到更相关的证据上。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["mcp-integrations","typed-decisions"],"stars":2,"forks":0,"openIssues":1,"license":"MIT","lastCommitAt":"2026-09-20T21:58:59Z","createdAt":"2026-09-19T00:51:00Z","metadataFetchedAt":"2026-09-21T22:14:21.155Z","avatarUrl":"https://avatars.githubusercontent.com/u/258440358?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/kindintelligence/jev-rust-review/blob/c6e116af0c5dfdcf039bf42d6ee4274ce1dbb28f/src/jev.rs","note":"自动发现的 Jev 集成证据"}]},{"id":"kamilpostrozny:pi-fast-jev-compaction","name":"pi-fast-jev-compaction","author":"KamilPostrozny","url":"https://github.com/KamilPostrozny/pi-fast-jev-compaction","category":"Context GC & Filter","plainSummary":"该扩展使用 TypeSafe Jev 逐个工具调用决定旧结果在模型 Context 中的保留，不改写用户与助手正文且保持持久会话记录完整。","plainSummaryEn":"Fast JEV compaction extension for pi","jevDecisionPoint":"Jev 逐个工具调用决定保留完整调用、仅删除结果还是完全删除调用。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"减少后续处理的冗余信息。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["context-compaction","classification-ranking"],"stars":2,"forks":0,"openIssues":1,"license":"MIT","lastCommitAt":"2026-09-20T18:05:51Z","createdAt":"2026-09-18T06:43:15Z","metadataFetchedAt":"2026-09-21T22:14:24.139Z","avatarUrl":"https://avatars.githubusercontent.com/u/27289342?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/KamilPostrozny/pi-fast-jev-compaction/blob/3cbdf934c236f9a57e60db12f343dcbf3904573f/extensions/fast-jev-core.ts","note":"自动发现的 Jev 集成证据"}]},{"id":"dougsong:jev-android","name":"jev-android","author":"dougsong","url":"https://github.com/dougsong/jev-android","category":"SDK & Decision Frameworks","plainSummary":"这是一个 Kotlin Android UI 自动化 SDK，由 TypeSafe Jev 或 DeepSeek 从当前屏幕控件中选择操作并通过无障碍服务执行，并附带示例应用。","plainSummaryEn":"A Kotlin Android SDK for UI automation powered by TypeSafe Jev, with an accessibility runtime and sample app.","jevDecisionPoint":"Jev根据当前屏幕可操作控件、历史动作效果和进度上下文选择下一步操作及其目标控件。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"在现有程序中复用接入代码，减少重复处理接口细节。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["multilanguage-sdk","typed-decisions"],"stars":2,"forks":0,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-20T17:41:56Z","createdAt":"2026-09-20T14:24:25Z","metadataFetchedAt":"2026-09-21T22:14:27.101Z","avatarUrl":"https://avatars.githubusercontent.com/u/39510437?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/dougsong/jev-android/blob/be8364ad1a7596172f8cd79c70d2462f4dd0f4dc/sdk/src/main/kotlin/io/github/jevandroid/JevProvider.kt","note":"自动发现的 Jev 集成证据"}]},{"id":"zurfyx:jev-browser-skill","name":"jev-browser-skill","author":"zurfyx","url":"https://github.com/zurfyx/jev-browser-skill","category":"Browser & OS Action","plainSummary":"这是一个让 Jev 驱动浏览器执行点击、输入和选择操作以完成指定目标的 Agent 技能，适用于 Claude Code 和 Codex。","plainSummaryEn":"Let Jev, TypeSafe's ~100ms decision model, drive your browser. A plug-and-play skill for Claude Code and Codex.","jevDecisionPoint":"Jev在每一步从候选操作和页面元素中选择要执行的浏览器动作及其目标元素。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"把选择和打分接进现有程序；暂无可核验的性能对照。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["browser-automation","typed-decisions"],"stars":1,"forks":0,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-21T19:56:24Z","createdAt":"2026-09-19T22:05:25Z","metadataFetchedAt":"2026-09-21T22:14:30.127Z","avatarUrl":"https://avatars.githubusercontent.com/u/193447?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/zurfyx/jev-browser-skill/blob/d5427f113db72ccf3b81119578080a89af31687e/api/jev.js","note":"自动发现的 Jev 集成证据"}]},{"id":"litshing:jevcore","name":"jevcore","author":"litshing","url":"https://github.com/litshing/jevcore","category":"SDK & Decision Frameworks","plainSummary":"JEV core 是向 Jev 发起有界批量判断请求的纯标准库客户端，包含成本守卫、缓存、校准、Harness 与命令行工具。","plainSummaryEn":"A judgement primitive for TypeSafe **System One / Jev** — ask N things × K typed\nquestions in bounded, cheap, fail-open requests — plus the **JEV harness**, the closed\nboundary in code that makes a Jev answer safe to consume. Standard library only.","jevDecisionPoint":"Jev 对每个条目回答类型化问题（如是否过时可丢弃）的 noul/score/choice 判断。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"增加一组可记录、可对照的判断信号。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["multilanguage-sdk","typed-decisions"],"stars":1,"forks":0,"openIssues":0,"license":null,"lastCommitAt":"2026-09-20T16:36:27Z","createdAt":"2026-09-20T16:33:34Z","metadataFetchedAt":"2026-09-21T22:14:33.152Z","avatarUrl":"https://avatars.githubusercontent.com/u/137675692?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/litshing/jevcore/blob/25fdfea90d87cc164434020567bf468c8eaeeb32/jevcore/cli.py","note":"自动发现的 Jev 集成证据"}]},{"id":"laguagu:jev-skills","name":"jev-skills","author":"laguagu","url":"https://github.com/laguagu/jev-skills","category":"SDK & Decision Frameworks","plainSummary":"该仓库为使用 Jev 构建应用的 Agent 提供实用技能与示例，涵盖 API 设置、决策模式、路由、排序与证据检查。","plainSummaryEn":"Practical agent skills and examples for building with Jev. API setup, routing, ranking, and evidence checks.","jevDecisionPoint":"对每个声明与证据段落对判定 supports/contradicts/irrelevant，再合成为 supported/contradicted/conflicting/not_stated/review。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"在现有程序中复用接入代码，减少重复处理接口细节。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["multilanguage-sdk","mcp-integrations"],"stars":1,"forks":0,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-21T14:20:18Z","createdAt":"2026-09-20T10:16:37Z","metadataFetchedAt":"2026-09-21T22:14:36.129Z","avatarUrl":"https://avatars.githubusercontent.com/u/49702497?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/laguagu/jev-skills/blob/c66fb6ea5404439572dc4049537ffb25c2f09a7a/examples/evidence/check.py","note":"自动发现的 Jev 集成证据"}]},{"id":"mattioofr:mcp-server-jev","name":"mcp-server-jev","author":"MattiooFR","url":"https://github.com/MattiooFR/mcp-server-jev","category":"MCP & Integrations","plainSummary":"该项目为 Codex、Claude 等 MCP 客户端提供 jev_evaluate 工具，使用 Jev 对提供的文本进行分类、判断和打分。","plainSummaryEn":"Typed AI decisions for Codex, Claude and any MCP client, powered by TypeSafe Jev. Classify, score and evaluate with one generic tool.","jevDecisionPoint":"对调用方提供的 state 内容执行 noul 二值判断、choice 分类选择和 score 等级评分的类型化决策。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"让现有 Agent 通过通用接口使用 Jev。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["mcp-integrations","typed-decisions"],"stars":1,"forks":0,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-20T16:15:45Z","createdAt":"2026-09-20T13:08:20Z","metadataFetchedAt":"2026-09-21T22:14:39.214Z","avatarUrl":"https://avatars.githubusercontent.com/u/13595614?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/MattiooFR/mcp-server-jev/blob/792bed11d1d2f987c8fcebd52c65b7859daf117c/src/client.mjs","note":"自动发现的 Jev 集成证据"}]},{"id":"suidouble:let-jev-speak","name":"let-jev-speak","author":"suidouble","url":"https://github.com/suidouble/let-jev-speak","category":"Creative Tools","plainSummary":"该项目通过一次一词地反复调用分类接口，让只做分类的 Jev 拼出自由文本回答。","plainSummaryEn":"TypeSafe's `/v1/systemone` endpoint classifies text — it returns a `choice`, a\n`score`, or a probability. It does not generate prose. This library makes it\ngenerate prose anyway: every word of the answer is a separate `choice` question\nover a vocabulary, and the loop feeds its own output back in as the prefix.","jevDecisionPoint":"Jev 通过 choice 问答逐词决定答案的下一个单词，并通过路由 choice 决定使用哪个领域词表。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"把语义判断接到已有的命令行工作流。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["creative-multimedia","typed-decisions"],"stars":1,"forks":0,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-20T15:49:17Z","createdAt":"2026-09-20T02:51:44Z","metadataFetchedAt":"2026-09-21T22:14:42.146Z","avatarUrl":"https://avatars.githubusercontent.com/u/135076785?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/suidouble/let-jev-speak/blob/393a4ca27333361e9a272972adc8d59236ca34cf/tg_bot/lib/typesafe.js","note":"自动发现的 Jev 集成证据"}]},{"id":"wjw66:deepseek-harness-jev-pre-compaction","name":"deepseek-harness-jev-pre-compaction","author":"wjw66","url":"https://github.com/wjw66/deepseek-harness-jev-pre-compaction","category":"Context GC & Filter","plainSummary":"在上下文接近deepseek-harness插件，在上下文压缩阈值前，自动识别并移除低价值的工具结果，降低上下文占用","plainSummaryEn":"A pre-compaction advisor for DeepSeek Harness. Runs before the standard `compaction-basic` backend, using TypeSafe JEV to safely prune low-value tool results from model context. Original session events stay in the append-only log; only the model-visible view is replaced with compact markers or archive pointers to reduce context bloat.","jevDecisionPoint":"判断当前是否应该提前裁剪上下文，以及每个工具结果是否值得继续保留。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"减少后续处理的冗余信息。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["context-compaction","typed-decisions"],"stars":0,"forks":0,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-21T17:57:13Z","createdAt":"2026-09-21T16:19:50Z","metadataFetchedAt":"2026-09-21T22:14:45.144Z","avatarUrl":"https://avatars.githubusercontent.com/u/49147577?v=4","summarySource":"readme-extractive","claimStatus":"优先保留投稿者与仓库原文，缺失语言自动补充；自动检查仅确认 Jev 集成证据，未经本站运行或性能复测。","claimStatusEn":"Author and repository text is preserved where clear; missing languages are enriched automatically. Checks establish source-level Jev integration, not runtime, safety, or performance validation.","evidence":[{"url":"https://github.com/wjw66/deepseek-harness-jev-pre-compaction/blob/b0b37f03fcae0c3b0b4116a9a9c3385588372d5c/README.md","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/wjw66/deepseek-harness-jev-pre-compaction/blob/b0b37f03fcae0c3b0b4116a9a9c3385588372d5c/README.zh.md","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/wjw66/deepseek-harness-jev-pre-compaction/blob/b0b37f03fcae0c3b0b4116a9a9c3385588372d5c/src/jev/client.ts","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/wjw66/deepseek-harness-jev-pre-compaction/blob/b0b37f03fcae0c3b0b4116a9a9c3385588372d5c/src/decision/apply.ts","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/wjw66/deepseek-harness-jev-pre-compaction/blob/b0b37f03fcae0c3b0b4116a9a9c3385588372d5c/src/trigger/gate.ts","note":"固定版本的 Jev 集成与说明来源"},{"url":"https://github.com/wjw66/deepseek-harness-jev-pre-compaction/blob/b0b37f03fcae0c3b0b4116a9a9c3385588372d5c/src/jev/protocol.ts","note":"固定版本的 Jev 集成与说明来源"}]},{"id":"sunil-sadasivan:jevernetes","name":"jevernetes","author":"sunil-sadasivan","url":"https://github.com/sunil-sadasivan/jevernetes","category":"Domain & Vertical Tools","plainSummary":"Jevernetes 在终端和本地仪表盘中实时采集 Kubernetes 日志，支持用 Jev 做语义搜索和分析、查看上下文，并把选中的日志证据整理后交给 coding Agent。","plainSummaryEn":"Live Kubernetes log analysis, contextual investigation, and agent handoff powered by Jev.","jevDecisionPoint":"Jev 对每条日志组判断是否值得调查的重要性、严重度和类别，以及是否匹配用户搜索问题。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"把语义判断接进已有的数据查询流程。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["domain-workflows","typed-decisions"],"stars":9,"forks":0,"openIssues":0,"license":"Apache-2.0","lastCommitAt":"2026-09-21T22:00:06Z","createdAt":"2026-09-20T23:13:05Z","metadataFetchedAt":"2026-09-21T22:16:29.807Z","avatarUrl":"https://avatars.githubusercontent.com/u/31223104?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/sunil-sadasivan/jevernetes/blob/1da673ac2eee851cd067e4669a42dad552eb7184/jevernetes/jev.py","note":"自动发现的 Jev 集成证据"}]},{"id":"ktaletsk:jevframe","name":"jevframe","author":"ktaletsk","url":"https://github.com/ktaletsk/jevframe","category":"Data & Search","plainSummary":"该库使用 TypeSafe Jev 对 pandas 和 Polars DataFrame 的每一行提出自然语言问题，并返回带完整概率分布的文本分类、情感分析和评分结果。","plainSummaryEn":"Semantic AI for pandas and Polars: classify text, analyze sentiment, and score DataFrame rows with natural-language questions and full probabilities using TypeSafe Jev.","jevDecisionPoint":"Jev 针对每一行上下文回答自然语言问题，判定情感、主题分类与等级评分并返回完整概率分布。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"把语义判断接进已有的数据查询流程。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["search-retrieval","typed-decisions"],"stars":7,"forks":1,"openIssues":1,"license":"MIT","lastCommitAt":"2026-09-21T18:52:47Z","createdAt":"2026-09-18T17:06:56Z","metadataFetchedAt":"2026-09-21T22:17:05.751Z","avatarUrl":"https://avatars.githubusercontent.com/u/8834829?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/ktaletsk/jevframe/blob/16bd3eae69b71284a75747fb4bf199b14ba1b075/src/jevframe/_engine.py","note":"自动发现的 Jev 集成证据"}]},{"id":"austinaway:working-memory-jev","name":"Working-Memory-Jev","author":"AustinAWay","url":"https://github.com/AustinAWay/Working-Memory-Jev","category":"Domain & Vertical Tools","plainSummary":"这是一个在本地运行的教学辅助工具，它使用 Jev API 分析英语教学文本在各阅读步骤中的主动分组、来源连接和需求变化，并与1–5槽位预算进行比较。","plainSummaryEn":"Edit `.env` locally and replace the placeholder with your own key. Keep this file private; Git ignores it. Do not paste the key into frontend code, a browser field, an issue, or a command that will be saved in shell history.","jevDecisionPoint":"Jev 对文本熟悉度、激活/背景状态、分组和关系归属做出有界选择，以估计每一步的记忆负荷。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"把选择和打分接进现有程序；暂无可核验的性能对照。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["domain-workflows","typed-decisions"],"stars":8,"forks":1,"openIssues":0,"license":null,"lastCommitAt":"2026-09-21T19:54:46Z","createdAt":"2026-09-18T15:23:03Z","metadataFetchedAt":"2026-09-21T22:17:55.400Z","avatarUrl":"https://avatars.githubusercontent.com/u/226801641?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/AustinAWay/Working-Memory-Jev/blob/cee3edb9922115d759778434878f6c56b483b390/backend/jev.py","note":"自动发现的 Jev 集成证据"}]},{"id":"endomorphosis:jevops","name":"JevOps","author":"endomorphosis","url":"https://github.com/endomorphosis/JevOps","category":"SDK & Decision Frameworks","plainSummary":"JevOps 是一个 TypeSafe / Jev 内核，提供门控、缓存、任务网格和 Lean IR 相关模块，但本身不编写 Lean 代码。","plainSummaryEn":"Jev is a **gate**, not a generator. This package does **not** write Lean.\nLake (or another oracle) lives in the implementation that *uses* the kernel.","jevDecisionPoint":"Jev 作为门控对候选方案执行 choice/score/noul 结构化评估，以决定是否放行或选择最优分支。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"把语义判断接到已有的命令行工作流。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["multilanguage-sdk","typed-decisions"],"stars":5,"forks":1,"openIssues":0,"license":"AGPL-3.0","lastCommitAt":"2026-09-21T20:50:18Z","createdAt":"2026-09-18T21:30:32Z","metadataFetchedAt":"2026-09-21T22:18:27.365Z","avatarUrl":"https://avatars.githubusercontent.com/u/3405202?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/endomorphosis/JevOps/blob/e085f65c179a9fd32259b3144d12a70d831b4f3e/jevops/typesafe_inference.py","note":"自动发现的 Jev 集成证据"}]},{"id":"andududu:jeview","name":"jeview","author":"andududu","url":"https://github.com/andududu/jeview","category":"SDK & Decision Frameworks","plainSummary":"Jeview 是一个本地网关，把代码发往 Jev 的每次调用转发出去，存入本地 SQLite 数据库，并在实时地图上展示。","plainSummaryEn":"An unofficial local visualizer for Jev (TypeSafe): a live view of every call your code makes. Not affiliated with TypeSafe AI.","jevDecisionPoint":"Jeview 自身不做决策，所有经网关转发的 System One 请求均由 Jev 负责回答和裁决。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"把选择和打分接进现有程序；暂无可核验的性能对照。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["multilanguage-sdk","database-vector"],"stars":7,"forks":2,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-21T20:32:53Z","createdAt":"2026-09-21T14:03:13Z","metadataFetchedAt":"2026-09-21T22:19:09.130Z","avatarUrl":"https://avatars.githubusercontent.com/u/129006005?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/andududu/jeview/blob/495a4e43e9d67e495a66ab0d9e55ac1c5c4040d6/src/jeview.ts","note":"自动发现的 Jev 集成证据"}]},{"id":"yijunyu:jev-rs","name":"jev-rs","author":"yijunyu","url":"https://github.com/yijunyu/jev-rs","category":"SDK & Decision Frameworks","plainSummary":"这是一个 Rust 编写的 Jev 兼容引擎，可在一次 prefill 中让任意 LLM 返回关于一段 state 的 noul、choice 和 score 类型判断及其概率，并可作为 MCP 工具供 Agent 调用。","plainSummaryEn":"System One judgments (noul/choice/score) from any LLM in one prefill — a Rust, Jev-compatible /v1/systemone engine","jevDecisionPoint":"Jev 根据输入的 state 文本对 noul、choice、score 等类型化问题给出带概率的判断。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"让现有 Agent 通过通用接口使用 Jev。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["multilanguage-sdk","typed-decisions"],"stars":4,"forks":0,"openIssues":0,"license":"Apache-2.0","lastCommitAt":"2026-09-21T21:45:49Z","createdAt":"2026-09-21T04:08:39Z","metadataFetchedAt":"2026-09-21T22:20:13.807Z","avatarUrl":"https://avatars.githubusercontent.com/u/884504?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/yijunyu/jev-rs/blob/8e29ca4a9ea6c926584b759f81b6d908458a4d7e/src/backend/typesafe.rs","note":"自动发现的 Jev 集成证据"}]},{"id":"choxos:jevchess","name":"jevchess","author":"choxos","url":"https://github.com/choxos/jevchess","category":"Domain & Vertical Tools","plainSummary":"该项目是让Jev与OpenRouter大模型、Stockfish引擎或人类实时对弈、回放并统计战绩的单页国际象棋应用。","plainSummaryEn":"Jev, TypeSafe's System One model, plays chess against any OpenRouter LLM, Stockfish and you. One-page web app with live moves, Jev's move probabilities, saved games and win rates.","jevDecisionPoint":"Jev在每一手从全部合法走法中以Choice选出最佳着法并以Score给出局面评估。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"把选择和打分接进现有程序；暂无可核验的性能对照。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["domain-workflows","typed-decisions"],"stars":2,"forks":0,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-20T04:42:58Z","createdAt":"2026-09-19T17:39:40Z","metadataFetchedAt":"2026-09-21T22:21:39.916Z","avatarUrl":"https://avatars.githubusercontent.com/u/6255852?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/choxos/jevchess/blob/602190f963e9b04def959b8519e6a2f2513590e0/server.mjs","note":"自动发现的 Jev 集成证据"}]},{"id":"kiwi0719:jev-edge","name":"jev-edge","author":"kiwi0719","url":"https://github.com/kiwi0719/jev-edge","category":"Security & Guardrails","plainSummary":"该项目是在 nginx/OpenResty 等网关入口处运行的三层请求过滤器，用 TypeSafe Jev 判断提示注入和滥用，并具备 fail-open、缓存和热更新能力。","plainSummaryEn":"Typed-judgment admission control at the traffic edge: three-layer prompt-injection and abuse filter for nginx/OpenResty, powered by TypeSafe Jev. Fail-open, cached, hot-reloadable.","jevDecisionPoint":"Jev 对网关入站请求文本是否为提示注入或滥用进行打分，返回概率以决定放行、拦截或旁路观察。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"给现有流程增加可检查的判断环节。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["security-guardrails","typed-decisions"],"stars":2,"forks":0,"openIssues":0,"license":"Apache-2.0","lastCommitAt":"2026-09-21T18:59:52Z","createdAt":"2026-09-21T13:46:04Z","metadataFetchedAt":"2026-09-21T22:22:27.417Z","avatarUrl":"https://avatars.githubusercontent.com/u/48856836?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/kiwi0719/jev-edge/blob/df85aec29e7f38279a7324f50f8f9413a2c4bc27/adapters/openresty/lib/resty/jev/providers/jev.lua","note":"自动发现的 Jev 集成证据"}]},{"id":"thehan-co:jevriel","name":"jevriel","author":"thehan-co","url":"https://github.com/thehan-co/jevriel","category":"SDK & Decision Frameworks","plainSummary":"这是一个帮助 AI Agent 使用 TypeSafe JEV 构建决策点、升级现有 LLM 工作流并测量效果的 skill 和插件。","plainSummaryEn":"**Codex · Claude Code · Portable skill** | [Apache-2.0 code and docs](LICENSE) | Early release","jevDecisionPoint":"JEV根据传入的状态对路由选择、相关性排序、候选抽取和证据支持度等有界问题返回类型化选择与评分以决定下一步。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"增加一组可记录、可对照的判断信号。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["multilanguage-sdk","typed-decisions","evaluation-benchmarks"],"stars":2,"forks":0,"openIssues":0,"license":"Apache-2.0","lastCommitAt":"2026-09-21T18:19:47Z","createdAt":"2026-09-21T16:12:04Z","metadataFetchedAt":"2026-09-21T22:23:02.793Z","avatarUrl":"https://avatars.githubusercontent.com/u/259903163?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/thehan-co/jevriel/blob/5896e01c466297371c8079648a43c6ac079da41f/server/providers.mjs","note":"自动发现的 Jev 集成证据"}]},{"id":"integrate-your-mind:jev-codex-plugin","name":"jev-codex-plugin","author":"integrate-your-mind","url":"https://github.com/integrate-your-mind/jev-codex-plugin","category":"MCP & Integrations","plainSummary":"该 Codex 插件提供 Jev 决策咨询、失败命令诊断和基于证据的完成情况检查功能。","plainSummaryEn":"Open-source Codex plugin for TypeSafe Jev decision consultation, failure diagnosis, and evidence-based completion review","jevDecisionPoint":"Jev 比较可用候选项并选出最优工具/模型/策略，或对失败原因与完成证据给出分类评估。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"让现有 Agent 通过通用接口使用 Jev。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["mcp-integrations","coding-agents","typed-decisions"],"stars":2,"forks":0,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-18T20:15:46Z","createdAt":"2026-09-18T15:09:14Z","metadataFetchedAt":"2026-09-21T22:24:34.921Z","avatarUrl":"https://avatars.githubusercontent.com/u/43152353?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/integrate-your-mind/jev-codex-plugin/blob/f57353ba6bc6288751d5a229e1359194d1b06a4c/source/jev-workflows/src/service.ts","note":"自动发现的 Jev 集成证据"}]},{"id":"rajivkuriakose:typesafe-jev-examples","name":"typesafe-jev-examples","author":"rajivkuriakose","url":"https://github.com/rajivkuriakose/typesafe-jev-examples","category":"SDK & Decision Frameworks","plainSummary":"该仓库提供通过 OpenRouter 调用 Jev 完成工单分类和文章重排序的可运行示例。","plainSummaryEn":"Worked examples for TypeSafe's Jev System One decision model, runnable today through OpenRouter","jevDecisionPoint":"Jev 对工单应分诊到哪个部门、是否紧急及业务影响等作类型化判断，并对候选文章是否真正回答用户问题作相关性判断。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"在现有程序中复用接入代码，减少重复处理接口细节。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["multilanguage-sdk","search-retrieval"],"stars":2,"forks":1,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-18T16:49:39Z","createdAt":"2026-09-18T15:11:04Z","metadataFetchedAt":"2026-09-21T22:25:09.549Z","avatarUrl":"https://avatars.githubusercontent.com/u/6995315?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/rajivkuriakose/typesafe-jev-examples/blob/824972389943a5bf1a351c0f0c846e43868e542f/src/jevx/client.py","note":"自动发现的 Jev 集成证据"}]},{"id":"aidil2105:jev-browser-pilot","name":"jev-browser-pilot","author":"aidil2105","url":"https://github.com/aidil2105/jev-browser-pilot","category":"Browser & OS Action","plainSummary":"该项目提供有界决策层，由代码负责观察、执行与验证，Jev 模型每步只选择一个候选操作。","plainSummaryEn":"A bounded decision layer for browser and desktop automation: a decision-only model picks one next step; the code owns perception, content, actuation and verification.","jevDecisionPoint":"Jev 根据当前页面观察文本和目标，从候选操作ID列表中选择下一步应执行的唯一操作。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"增加一组可记录、可对照的判断信号。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["browser-automation","typed-decisions"],"stars":2,"forks":1,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-18T11:54:09Z","createdAt":"2026-09-18T06:13:17Z","metadataFetchedAt":"2026-09-21T22:25:33.265Z","avatarUrl":"https://avatars.githubusercontent.com/u/134305645?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/aidil2105/jev-browser-pilot/blob/a7bd4cebcf869fd063aba4f7eef4aa50c3618ef7/src/jev_pilot/providers/jev.py","note":"自动发现的 Jev 集成证据"}]},{"id":"yibie:laya-jev-lab","name":"laya-jev-lab","author":"yibie","url":"https://github.com/yibie/laya-jev-lab","category":"Routing & Cost Optimization","plainSummary":"该仓库对比 Jev 与 Laya 在中文客服工单分类上的测量结果，并提供本地优先的 cascade 实验脚本。","plainSummaryEn":"Independent measurements of typed-decision models: Jev (TypeSafe API) vs Laya (open weights), and a local-first cascade that matches Jev's accuracy at 1.8x the speed","jevDecisionPoint":"Jev 对每条中文客服消息执行 choice 分类决策，判断其所属类别。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"把语义判断接到已有的命令行工作流。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["llm-routing-cost","typed-decisions","evaluation-benchmarks"],"stars":2,"forks":1,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-20T06:00:10Z","createdAt":"2026-09-20T06:00:15Z","metadataFetchedAt":"2026-09-21T22:27:13.881Z","avatarUrl":"https://avatars.githubusercontent.com/u/365484?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/yibie/laya-jev-lab/blob/30ba64dceb7da058b306c104a571f1242dbd582b/cascade/run2-jev.mjs","note":"自动发现的 Jev 集成证据"}]},{"id":"gaborishka:jevtown","name":"jevtown","author":"gaborishka","url":"https://github.com/gaborishka/jevtown","category":"High-Frequency & Simulation","plainSummary":"把 Jev 的结构化判断接进程序；具体用途与决策流程请查看项目源码。","plainSummaryEn":"A check costs from half a cent (a text that dies in the first wave) to ten cents (one that reaches all 10,000), and takes from 3 seconds to a minute. The interface comes in Ukrainian and English, and so do the personas: a text is read by the crowd that speaks its language, 10,000 Ukrainians or 10,000 English speakers, so there is nothing to choose.","jevDecisionPoint":"Jev 为每个 AI 人格决定对帖子的具体反应（滑过、点赞、转发、屏蔽、购买等）及曝光与追问的概率。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"在连续交互中观察决策效果；频率依实际运行而定。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["games-simulation","typed-decisions"],"stars":2,"forks":0,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-21T14:39:36Z","createdAt":"2026-09-20T05:41:46Z","metadataFetchedAt":"2026-09-21T22:27:41.831Z","avatarUrl":"https://avatars.githubusercontent.com/u/25786922?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/gaborishka/jevtown/blob/f7cf208f3825cee48215ded684ddd39aa71ad20b/public/shared/jev.js","note":"自动发现的 Jev 集成证据"}]},{"id":"daniel4x:jevemon","name":"JevEmon","author":"daniel4x","url":"https://github.com/daniel4x/JevEmon","category":"Domain & Vertical Tools","plainSummary":"JevEmon从宝可梦火红ROM内存读取地图与队伍状态，让Jev在可到达目的地中选择行走目标并指挥野生宝可梦对战，从主角家一路走到常青市。","plainSummaryEn":"Verified milestones so far: Jev can leave the Player's House, cross Pallet Town, deliver itself through Route 1 (fighting and winning any wild encounters along the way), and reach Viridian City. Further legs of the journey (Oak's Parcel, the first Gym) aren't built yet — the walk currently ends the run once it reaches Viridian City.","jevDecisionPoint":"Jev从已验证可达的目的地列表中选择下一站行走目标，遇野怪时再选择出招或换宠。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"把选择和打分接进现有程序；暂无可核验的性能对照。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["domain-workflows","typed-decisions"],"stars":2,"forks":0,"openIssues":0,"license":"GPL-3.0","lastCommitAt":"2026-09-21T13:25:03Z","createdAt":"2026-09-21T12:50:05Z","metadataFetchedAt":"2026-09-21T22:28:13.761Z","avatarUrl":"https://avatars.githubusercontent.com/u/16572424?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/daniel4x/JevEmon/blob/572454c69bf791f766ece736ffa6d6c954c6a075/jevemon/decisions.py","note":"自动发现的 Jev 集成证据"}]},{"id":"jonymusky:jev-browser-qa","name":"jev-browser-qa","author":"jonymusky","url":"https://github.com/jonymusky/jev-browser-qa","category":"Browser & OS Action","plainSummary":"该项目用 Playwright 执行并录制浏览器流程，用 Jev 判断自然语言断言、按意图选择控件和执行目标导向点击循环，并提供 JSON-flow CLI、运行看板和 Agent skill。","plainSummaryEn":"Browser QA where Playwright drives and films, and TypeSafe Jev judges. JSON-flow CLI for agents, run dashboard, agent skill.","jevDecisionPoint":"Jev负责判断自然语言断言是否成立、按意图在可见控件中选择点击目标、以及在目标驱动循环中决定下一步点击。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"把语义判断接到已有的命令行工作流。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["browser-automation","typed-decisions"],"stars":2,"forks":0,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-19T20:39:15Z","createdAt":"2026-09-19T20:12:58Z","metadataFetchedAt":"2026-09-21T22:29:06.303Z","avatarUrl":"https://avatars.githubusercontent.com/u/414404?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/jonymusky/jev-browser-qa/blob/ca6d5a2d46623f9f8606aabc10f1143afd070090/src/jev/client.ts","note":"自动发现的 Jev 集成证据"}]},{"id":"newuser7171:jev-gamepilot","name":"jev-gamepilot","author":"newuser7171","url":"https://github.com/newuser7171/jev-gamepilot","category":"High-Frequency & Simulation","plainSummary":"该项目是基于 Laya 和 TypeSafe Jev System One 的自主游戏 Agent，可捕获游戏画面并在 Windows PC 游戏和 Android 手机游戏上执行操作。","plainSummaryEn":"**Jev-GamePilot** is a universal autonomous AI gaming agent powered by **Laya (local sub-30ms System One inference)** and **TypeSafe's Jev System One** (`Choice`, `Score`, `Noul`). It captures real-time gameplay at 60+ FPS, fuses instant local reflexes with high-level strategic reasoning, and executes physical hardware inputs across Windows PC games and connected Android phones.","jevDecisionPoint":"Jev根据障碍物类型与距离选择跳跃、低头或正常奔跑，并评估碰撞威胁等级与是否需要空中快速下落。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"把选择和打分接进现有程序；暂无可核验的性能对照。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["games-simulation","typed-decisions"],"stars":2,"forks":0,"openIssues":0,"license":null,"lastCommitAt":"2026-09-21T21:19:17Z","createdAt":"2026-09-19T09:29:21Z","metadataFetchedAt":"2026-09-21T22:29:35.795Z","avatarUrl":"https://avatars.githubusercontent.com/u/217369131?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/newuser7171/jev-gamepilot/blob/3f03d6ec180b92cf371f53e327d2cd5c18e9d4b4/jev_brain.py","note":"自动发现的 Jev 集成证据"}]},{"id":"soundblaster:jev4mellea","name":"Jev4Mellea","author":"SoundBlaster","url":"https://github.com/SoundBlaster/Jev4Mellea","category":"SDK & Decision Frameworks","plainSummary":"这是一个将 Jev 语义检查接入 Mellea 的 Python 适配器，用于验证文本、分类文本或按有序量表评分。","plainSummaryEn":"This is an unofficial, synchronous adapter. Jev evaluates text; it does not\ngenerate or repair it.","jevDecisionPoint":"Jev 负责对候选文本执行 Noul 验证、Choice 分类和 Score 评分等语义判定。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"增加一组可记录、可对照的判断信号。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["multilanguage-sdk","typed-decisions"],"stars":2,"forks":0,"openIssues":0,"license":"Apache-2.0","lastCommitAt":"2026-09-21T20:22:32Z","createdAt":"2026-09-17T13:12:43Z","metadataFetchedAt":"2026-09-21T22:29:57.913Z","avatarUrl":"https://avatars.githubusercontent.com/u/809319?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/SoundBlaster/Jev4Mellea/blob/c6b237271070af62dfffda6b9644595a29651984/src/mellea_jev/providers/typesafe.py","note":"自动发现的 Jev 集成证据"}]},{"id":"nourhelmi:pi-jev-compaction","name":"pi-jev-compaction","author":"nourhelmi","url":"https://github.com/nourhelmi/pi-jev-compaction","category":"Context GC & Filter","plainSummary":"该 Pi 扩展使用 Jev 评估较旧的工具输出并将其从上下文中清除，同时保留原始会话记录以便通过检索取回。","plainSummaryEn":"Automatic Jev context clearing for Pi. Keep the conversation, prune stale tool output, retrieve originals without rerunning commands.","jevDecisionPoint":"Jev 对最多16个陈旧工具输出打分，判断其仍被需要的概率，低于阈值则清除。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"减少后续处理的冗余信息。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["context-compaction","classification-ranking"],"stars":2,"forks":1,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-21T20:02:39Z","createdAt":"2026-09-18T16:28:37Z","metadataFetchedAt":"2026-09-21T22:30:37.318Z","avatarUrl":"https://avatars.githubusercontent.com/u/107139955?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/nourhelmi/pi-jev-compaction/blob/103c0b006f3316705a49501db22bfe8727ebcdd8/src/pruning.ts","note":"自动发现的 Jev 集成证据"}]},{"id":"codeia-academy:jev-mcp","name":"jev-mcp","author":"CodeIA-Academy","url":"https://github.com/CodeIA-Academy/jev-mcp","category":"MCP & Integrations","plainSummary":"该项目是一个无依赖的本地 MCP 服务器，将 Jev 作为 ask_jev 和 list_jev_models 工具提供给 Agent 使用。","plainSummaryEn":"MCP local que expone Jev (TypeSafe) como herramienta para Claude Code, Codex, Hermes y cualquier agente: ask_jev y list_jev_models, sin dependencias","jevDecisionPoint":"Jev 根据传入的 state 对 choice/score/noul 类型问题返回带概率和置信度的分类、评分或真值判断，供代理决定路由与后续动作。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"把语义判断接到已有的命令行工作流。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["mcp-integrations","typed-decisions"],"stars":1,"forks":0,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-20T17:47:36Z","createdAt":"2026-09-20T17:24:10Z","metadataFetchedAt":"2026-09-21T22:31:07.580Z","avatarUrl":"https://avatars.githubusercontent.com/u/243828955?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/CodeIA-Academy/jev-mcp/blob/3b51c2438eb5c8f404dc381c06b5e00e0ab71146/server.py","note":"自动发现的 Jev 集成证据"}]},{"id":"q93304989-bit:jev-lab","name":"jev-lab","author":"q93304989-bit","url":"https://github.com/q93304989-bit/jev-lab","category":"SDK & Decision Frameworks","plainSummary":"最简 Jev 调用演示器：单页分类器，把请求 JSON、概率分布、confidence、耗时与 token 都摊开给你看","plainSummaryEn":"This repository provides a single-page classifier demo that sends text with choice questions to Jev and displays the request JSON, probability distribution, confidence, latency, and token usage.","jevDecisionPoint":"Jev根据输入文本在自定义类别上做choice分类，给出各标签概率分布与confidence。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"把语义判断接到已有的命令行工作流。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["multilanguage-sdk","typed-decisions"],"stars":1,"forks":0,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-20T15:35:54Z","createdAt":"2026-09-20T15:36:54Z","metadataFetchedAt":"2026-09-21T22:31:40.512Z","avatarUrl":"https://avatars.githubusercontent.com/u/308038056?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/q93304989-bit/jev-lab/blob/8a6a84af140327f9edae24f6532531379ebe0296/server.py","note":"自动发现的 Jev 集成证据"}]},{"id":"leddoengano:jev-eyes","name":"jev-eyes","author":"LeddoEngano","url":"https://github.com/LeddoEngano/jev-eyes","category":"MCP & Integrations","plainSummary":"jev-eyes 在本地把图片转成 Jev 可决策的文字和布局状态，并提供 CLI、MCP server 和 Agent skill。","plainSummaryEn":"Also in the state: `image` (size, source), `blocks` (`[x, y, w, h]` boxes with OCR confidence) and, if installed, `labels`. `see(img, compact=True)` keeps only `image`, `text` and top label names when tokens matter more than positions.","jevDecisionPoint":"Jev基于OCR文本与空间布局构成的state判断图像内容属性以及是否需要升级到视觉模型。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"让现有 Agent 通过通用接口使用 Jev。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["mcp-integrations","typed-decisions"],"stars":1,"forks":0,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-20T15:20:54Z","createdAt":"2026-09-20T05:41:23Z","metadataFetchedAt":"2026-09-21T22:32:10.342Z","avatarUrl":"https://avatars.githubusercontent.com/u/49451237?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/LeddoEngano/jev-eyes/blob/9ae9a6dc2088359a52d7092891bbf7c8f2030936/benchmarks/jev_accuracy.py","note":"自动发现的 Jev 集成证据"}]},{"id":"ethanalgox:jev-trading","name":"jev-trading","author":"EthanAlgoX","url":"https://github.com/EthanAlgoX/jev-trading","category":"Domain & Vertical Tools","plainSummary":"1. Keep **Demo** selected (`演示体验`).\n2. Click **Try a decision** (`体验一次决策`).\n3. Review the action, probabilities, and data quality on the right.\n4. Open a record in **Recent analyses** (`最近分析`) or select **Export evidence** (`导出完整依据`).","plainSummaryEn":"**Jev Trading is a stock decision service that runs on your computer.** Use the web workbench or call it from your own software over HTTP. It sends market data, indicators, fundamentals, and news collected by AIStock to a model, then returns a structured decision with a record of the evidence.","jevDecisionPoint":"Jev 根据行情、基本面、新闻与持仓情况在买入/卖出/持有三个动作中做出分类决策并返回概率分布。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"增加一组可记录、可对照的判断信号。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["domain-workflows","typed-decisions"],"stars":1,"forks":0,"openIssues":0,"license":null,"lastCommitAt":"2026-09-20T14:55:53Z","createdAt":"2026-09-20T04:20:00Z","metadataFetchedAt":"2026-09-21T22:32:36.187Z","avatarUrl":"https://avatars.githubusercontent.com/u/22077167?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/EthanAlgoX/jev-trading/blob/779eda2c27315b051af637a7e33437a3cf72178a/src/jev_trading/model.py","note":"自动发现的 Jev 集成证据"}]},{"id":"reiswaffel78:jev-agent-toolkit","name":"jev-agent-toolkit","author":"reiswaffel78","url":"https://github.com/reiswaffel78/jev-agent-toolkit","category":"MCP & Integrations","plainSummary":"该仓库提供便携式 Agent Skill，教编码 Agent 将 Jev 用作有界判断层，并附带可选 MCP 桥及面向代码、浏览器研究、Blender 和 Unreal Engine 的工作流。","plainSummaryEn":"Jev-first portable Agent Skill and optional MCP bridge for Claude Code, Codex, Cursor and compatible agents.","jevDecisionPoint":"Jev 对工单的子系统归属、严重程度及退款与安全意图进行有界语义判断并返回校准概率。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"让现有 Agent 通过通用接口使用 Jev。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["mcp-integrations","typed-decisions"],"stars":1,"forks":0,"openIssues":1,"license":"MIT","lastCommitAt":"2026-09-20T14:47:01Z","createdAt":"2026-09-19T10:47:33Z","metadataFetchedAt":"2026-09-21T22:33:09.969Z","avatarUrl":"https://avatars.githubusercontent.com/u/135417706?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/reiswaffel78/jev-agent-toolkit/blob/700a3fdce348b7b96868898623f7de49dd0b426b/examples/direct-api/triage.py","note":"自动发现的 Jev 集成证据"}]},{"id":"sqaassl:openclaw-jev-compaction","name":"openclaw-jev-compaction","author":"SqaaSSL","url":"https://github.com/SqaaSSL/openclaw-jev-compaction","category":"Context GC & Filter","plainSummary":"这是一个为 OpenClaw 提供的逐字上下文压缩引擎，它请 Jev 判断哪些工具调用和工具结果仍需保留并删除其余内容，且从不做总结。","plainSummaryEn":"Verbatim context compaction for OpenClaw: a context engine powered by TypeSafe's Jev. Drops stale tool calls and results, never summarizes.","jevDecisionPoint":"Jev 对每个工具调用及其结果分别打分，决定原文保留、截断结果还是整体删除。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"减少后续处理的冗余信息。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["context-compaction","classification-ranking"],"stars":1,"forks":0,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-20T11:39:21Z","createdAt":"2026-09-20T07:29:25Z","metadataFetchedAt":"2026-09-21T22:34:05.501Z","avatarUrl":"https://avatars.githubusercontent.com/u/160170823?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/SqaaSSL/openclaw-jev-compaction/blob/6a2958ed4ed343cb15457bece8d357e6e6414912/src/core/request.ts","note":"自动发现的 Jev 集成证据"}]},{"id":"muse0509:jev-preflight","name":"jev-preflight","author":"muse0509","url":"https://github.com/muse0509/jev-preflight","category":"Security & Guardrails","plainSummary":"jev-preflight：Jev 对行为回归、鉴权等八个风险轴一次性打分，判断本轮变更是否达到需复查的高风险阈值。","plainSummaryEn":"A bounded Jev risk check for Claude Code: eight risk axes, one request, one optional reinspection.","jevDecisionPoint":"Jev 对行为回归、鉴权等八个风险轴一次性打分，判断本轮变更是否达到需复查的高风险阈值。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"把语义判断接到已有的命令行工作流。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["security-guardrails","typed-decisions"],"stars":1,"forks":0,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-20T09:31:19Z","createdAt":"2026-09-18T13:54:37Z","metadataFetchedAt":"2026-09-21T22:34:56.470Z","avatarUrl":"https://avatars.githubusercontent.com/u/184503642?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/muse0509/jev-preflight/blob/c143b921253fd814bcd3c53d638a0cc8d390d49c/internal/jev/client.go","note":"自动发现的 Jev 集成证据"}]},{"id":"zerodegress:jevinf","name":"jevinf","author":"zerodegress","url":"https://github.com/zerodegress/jevinf","category":"SDK & Decision Frameworks","plainSummary":"Jev 这一系决策模型的推理引擎：每条候选路径按分段前向计算并复用前缀，上面架一层符合 Jev wire\n契约的服务。目前接上的后端是 NanoJev。","plainSummaryEn":"An inference engine for decision models of the Jev kind: each candidate path runs as segmented\nforwards with prefix reuse, and the Jev wire contract is served on top. NanoJev is the backend wired\nup today.","jevDecisionPoint":"Jev 根据事件状态对计费、紧急程度、负责团队和客户情绪等问题做出是否判断、选项选择和程度打分。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"把语义判断接到已有的命令行工作流。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["multilanguage-sdk","typed-decisions"],"stars":1,"forks":0,"openIssues":1,"license":"MIT","lastCommitAt":"2026-09-20T05:25:21Z","createdAt":"2026-09-19T13:04:55Z","metadataFetchedAt":"2026-09-21T22:36:36.746Z","avatarUrl":"https://avatars.githubusercontent.com/u/60560055?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/zerodegress/jevinf/blob/4c9827aeab1e22ca4519158dea2964176d8ba310/scripts/jev_conformance.py","note":"自动发现的 Jev 集成证据"}]},{"id":"wenchenxi:jev-console","name":"jev-console","author":"wenchenxi","url":"https://github.com/wenchenxi/jev-console","category":"CLI & Pipelines","plainSummary":"给 TypeSafe 的 **Jev（System One 决策模型）** 做的本地小控制台 + 命令行。","plainSummaryEn":"Jev does not generate text. You send it a **state** plus a set of **typed questions**, and it answers each one with a typed value and a calibrated probability:","jevDecisionPoint":"Jev 根据用户提供的 state 对 noul 是否判断、choice 多选一、score 档位打分问题给出校准概率和置信度。","jevDecisionPointEn":"Jev returns a structured decision for the local program; consult the source for the exact decision policy.","highlightBenefit":"把语义判断接到已有的命令行工作流。","highlightBenefitEn":"Adds structured choices or scores to the workflow; performance and cost benefits have not been independently verified.","tags":["cli-git-gates","typed-decisions"],"stars":1,"forks":0,"openIssues":0,"license":"MIT","lastCommitAt":"2026-09-20T03:52:16Z","createdAt":"2026-09-20T02:03:11Z","metadataFetchedAt":"2026-09-21T22:37:46.934Z","avatarUrl":"https://avatars.githubusercontent.com/u/78245923?v=4","summarySource":"readme-extractive","claimStatus":"根据仓库简介与 README 自动提炼；决策机制为规则归类，待人工复核，未独立测试性能。","claimStatusEn":"Based on repository metadata and README with rule-based classification; pending human review, with no independent runtime or performance verification.","evidence":[{"url":"https://github.com/wenchenxi/jev-console/blob/8f9c17d570346b982d6d590ad148bd0f1a063b65/jev_cli.py","note":"自动发现的 Jev 集成证据"}]}]