ai_claims
Use as the AI/LLM example step — turn an input's `source` values into `claims`, recording a trace entry per value so a trace_gate can block anything not traceable to the input (the no-fabrication pattern). Stand-in for a real LLM call. Single tool.
- License 许可
- MIT
- 最近更新
- 2026-10-30
- Stars
- ★ 211
- Forks
- 21
01 前言 Front Matter
- name
- ai_claims
- description
- Use as the AI/LLM example step — turn an input's `source` values into `claims`, recording a trace entry per value so a trace_gate can block anything not traceable to the input (the no-fabrication pattern). Stand-in for a real LLM call. Single tool.
02 资源文件 1
- 脚本tool.py1005 B
03 SKILL.md 正文
Example skill — ai_claims(AI 範例:每個值都可溯源)
workflow/examples/ai.json 的唯一一步。示範「模型產出的每個值都要能溯回輸入」:
本 stand-in 只把輸入的 source 原樣抄成 claims,並為每個值記一筆 trace。
綁定的單一工具
python3(標準庫)。換成真的 LLM 呼叫時,保留下游的trace_gate——模型自己發明、輸入裡找不到的值會被擋下。
依賴(完整宣告;缺項明確報錯)
python>= 3.8
參數化(無硬編碼)
- 輸入:
--in <artifact.json>(讀data.source,一個清單) - 輸出:
--out <claims artifact.json>
介面
- 輸出 artifact:
{"schema":"claims@1","produced_by":"ai_claims","data":{"claims":[...]},"trace":[{"value":"<str>","source":"<輸入檔名>","locator":"source[i]"}]} - 閘門:
trace_gate,args.fields = "claims"。
執行
python3 skills/ai_claims/tool.py --in example/inputs/ai_source.json --out <out/ai.json>
04 同仓库更多 Skills
second-opinion
Use when asked to get a second opinion on a finished DRAFT, fact-check or double-check a report's claims against its evidence, sanity-check before sign-off, or red-team an output for fabrication / overreach / invented IDs — even without the word "honesty". Works on an agentic-sop-kit run dir (FULL, uses the run's trace) or any plain document (DEGRADED, you supply the inputs). | 要對一份已完成的 DRAFT 取得「第二意見」、在核准前查核報告主張是否有輸入佐證、抓臆造/灌水/過度宣稱/捏造編號時使用;吃 agentic-sop-kit 的 run 目錄(FULL,用 trace)或任何純文件(DEGRADED,需自備輸入)。不適用:稽核 agent 架構有沒有拆好/是不是 mega-agent(改用 agentic-workflow-audit);審查原始碼品質/bug/資安(改用 /code-review)。本 skill 只查「這次產出的主張 vs. 證據」誠實度,只產 DRAFT,核准永遠人擁有。
plugin-forge
Use when creating, scaffolding, or linting a Claude Code plugin — generate a grammar-conformant plugin skeleton, or validate a plugin / whole marketplace against house invariants (manifest, frontmatter, stdlib-only, hook protocol, test harness). | 要新建 / scaffold 一個 Claude Code plugin,或 lint 既有 plugin / 整個 marketplace 是否符合房規時使用。
good-plugin
demo good plugin
bad-plugin
暂无简介
alcoa-guard
Use when checking a record / dataset / GMP document for ALCOA+ data integrity — attribution, contemporaneity (backdating), completeness, accuracy (recompute / in-spec), consistency — before a human approves. DEGRADED mode lints any CSV/JSON record against a declared .alcoa.json contract; FULL mode snaps onto an agentic-sop-kit run dir. | 要檢查紀錄/資料/GMP 文件的資料完整性(ALCOA+:可歸屬、同步性/防 backdating、完整、正確、一致)時使用。輸出為 DRAFT,人擁有最終判定;確定性檢查只覆蓋可機械驗證的部分,其餘列為人判斷清單,絕不臆造。
sop-flow
Run the agentic-sop-kit workflow (extract→compute→report, or a declared topology) and report the DRAFT result. Works out-of-the-box (bundled demo) even before a project adopts the kit.