원클릭으로
documentation-governance
Use when auditing documentation structure, duplicate markdown, README governance, or docs/docs_readable boundaries.
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
메뉴
Use when auditing documentation structure, duplicate markdown, README governance, or docs/docs_readable boundaries.
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
SOC 직업 분류 기준
Adapter entry for bootstrapping agent rules, contracts, validation, task packages, reports, and handoff boundaries for financial data engineering or operations projects. Canonical source is in skills/financial-data-agent-bootstrap.
Use when bootstrapping agent rules, contracts, validation, task packages, reports, and handoff boundaries for financial data engineering or operations projects.
Use when evaluating, scoring, diagnosing, or explicitly refactoring a single skill under controlled governance.
Use when onboarding a new maintainer or AI agent, scanning an unfamiliar repository, or generating a takeover packet.
Use when preparing a handoff task package, bounded execution, or execution report across AI agents.
Use when assessing migration readiness for a financial-data Python project with script sprawl, Excel assets, or Wind/Desktop coupling.
| name | documentation-governance |
| description | Use when auditing documentation structure, duplicate markdown, README governance, or docs/docs_readable boundaries. |
| metadata | {"triggers":["audit repository documentation structure","detect duplicate markdown documents","fix README governance issues","classify docs and docs_readable layers","archive or merge outdated documentation"],"side_effects":["read_only","write_files"]} |
This execution-focused skill definition keeps the behavior, invocation shape, and adapter-facing contract unchanged while moving explanation-oriented content into supporting assets.
这个 pattern 的核心是把文档治理稳定组织为 audit -> report -> fix(optional):
--write、--report、--json 等输出控制参数report,包括治理摘要、高优先级问题、建议动作、可生成目标和冲突清单fix 仅在显式授权时发生,例如补写缺失 README 章节或生成报告文件;默认不自动进入写入模式Input:
--write、--report、--json 等输出控制参数Process:
auditreportfix(optional)Output:
reportskill_assessment_output, trimmed to the audit scenario这样组织的原因是,文档治理首先需要稳定判断“哪里失控、哪里重复、哪里越权”,而不是立刻改动文件。先 audit,再 report,最后才进入 fix,可以避免把治理规则误用成粗暴重写命令。
先审计文档事实,再做结构治理。
Audit documentation facts before enforcing structure changes.
一个主题只保留一个工程事实源。
Keep one engineering source of truth per topic.
可变项目状态事实必须留在已声明的 current-state SSOT 中。
Mutable project-status facts must stay in the declared current-state SSOT. README.md、docs/README.md、docs/technical/、CLAUDE.md、AGENTS.md 和 blueprint docs 可以引用 SSOT,但不应复制 active phase status、next-phase decisions、latest validation results、blocker status 或 pending-merge state,避免形成第二事实源。
默认不改变项目,只做观察和结构化输出。
Do not modify the project unless explicitly allowed.
docs_readable/ 只能是衍生层,不能反向成为权威层。
Treat docs_readable/ as a derivative layer, not an authority layer.
fix 必须显式触发,并且晚于审计结论。
Any fix action must be explicit and come after the audit conclusion.
audit:解析治理输入。
读取项目根目录、默认治理配置、项目覆盖配置和显式参数;确认 --config、--dry-run、--write、--report、--json 等输入,并识别 style guide、README.md、docs/、docs_readable/ 以及本次适用的分类与命名规则。
audit:扫描文档树并执行检查。
收集 markdown 文件,检查 heading 结构、README 必需章节、分类放置、层级归属、禁止命名、重复主题、单一事实源冲突、readable layer 第二事实源风险、归档候选与可读层生成目标。此阶段默认只观察,不写入文档。
report:输出治理结果。
生成工程层摘要、阅读层摘要、高优先级问题、建议动作、冲突清单和归档候选;如启用 --json,输出机器可读结果;如指定 --report,则将治理报告写入目标文件。报告应清楚区分“问题发现”和“建议修复”,而不是把建议动作当成已执行事实。需要 assessment 口径时,引用 shared assessment output protocol,并按文档审计场景裁剪字段,不强制所有字段满配。
fix:决定是否进行显式修复。
只有在用户明确授权,且治理结论已经清楚时,才允许进入 fix,例如使用 --write 为 README 补齐缺失章节,或在后续单独执行归档、去重、readable 生成等受控动作。默认不应自动合并、删除、重命名或改写文档内容。
输出治理回顾。
在最终结果中说明本次治理使用的配置来源、扫描范围、主要冲突、建议动作,以及哪些内容只是报告建议、哪些内容已经被显式写入或生成。
fix 必须显式触发;--write 仅用于治理已经清楚后的 README 章节补写,不应扩展为任意文档重写docs/ 应继续被视为工程事实层,docs_readable/ 只能是衍生层,不能升格为第二权威层SKILL.md 仍是该 skill 的顶层规则表达,不应被辅助文件覆盖human_ai_shared, human_primary_archive, ai_only_wrapper, unknown_or_mixed.document_layer model.