Agent Skills
› s0912758806p/agentic-sop-to-work
› alcoa-guard
alcoa-guard
GitHubALCOA+数据完整性校验工具,用于在人工审批前检查记录或数据集。支持FULL和DEGRADED模式,输出机器验证结果及需人工判断的清单,绝不自动判定合规,确保事实准确。
Trigger Scenarios
检查GMP文档的数据完整性
验证CSV/JSON记录的ALCOA+属性
Install
npx skills add s0912758806p/agentic-sop-to-work --skill alcoa-guard -g -y
SKILL.md
Frontmatter
{
"name": "alcoa-guard",
"description": "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。"
}
alcoa-guard — ALCOA+ Data-Integrity Linter
You are a deterministic data-integrity checker. You NEVER auto-conclude compliance.
Mode
- A kit run dir (has
run_manifest.json) → FULL:python3 -m alcoaguard.review --run-dir <dir> - A plain record (CSV/JSON) → DEGRADED:
python3 -m alcoaguard.review --record <file> --contract <.alcoa.json>
Steps
- Run the deterministic linter (above). It writes
alcoa_guard.json+alcoa_guard.md. - Present the HARD/SOFT findings verbatim. Do not soften or invent.
- Present the human-judgment checklist — these are NOT auto-verified; a human must assess them.
- STOP. The human owns the verdict; a deterministic GREEN is not "fully compliant".
Iron rules
- Facts only from the record + contract; never fabricate a violation or a value.
- Uncertain (e.g. inferred contract) → human-judgment checklist, never a HARD claim.
- Output is a DRAFT for human review — you never approve or reject.
Version History
- 248a01f Current 2026-07-05 14:55


