docs-stale-audit
Skill4 starsChanged 4 days agoMITView on GitHub
文档/技能改动后的墓碑描述审计闭环:audit-docs.mjs 的 --changed/--staged 用法、高置信与复查候选判级、docs-audit:allow 豁免、修复流程、git 钩子拦截。Triggers: 文档审计, 墓碑描述, 改动后审计, 过时描述, audit-docs, --changed, --staged, 复查候选, docs-audit:allow, 钩子拦截
In this repository15 other files it gives its agents.
- AGENTS.mdAGENTS.md
- dsh/AGENTS.mdAGENTS.md
- android-apk-download-verifySkill
- auto-skill-creatorSkill
- dsh-turn-failure-diagnosisSkill
- git-branch-consolidationSkill
- opencode-oauth-context-capSkill
- opencode-share-kitSkill
- phased-refactor-and-reportSkill
- plan-reviewSkill
- powershell-utf8-safetySkill
- quality-baselineSkill
- test-verificationSkill
- white-bear-safe-rulesSkill
- windows-openssh-serverSkill
The file
Files
文档墓碑描述审计(docs-current-state-only 闭环)
<role>
回答「改动后怎么确认文档只写现状」:跑审计 → 看判级 → 修文档 → 进完成报告。与 AGENTS.md 的 docs-current-state-only 规则配套,把写前引导延伸为写后检查。
</role>边界职责
auto-skill-creator:技能的创建/更新/同步/冲突检查流程。powershell-utf8-safety:写文件编码安全。- 本技能:文档正文的墓碑描述审计(命令、判级、豁免、修复)。
何时跑
- 功能/修复收尾、文档或技能改动后、提交前(完成门禁的一部分)。
- 项目内只覆盖本次改动:
node ~/.dsh/rules/audit-docs.mjs --changed - 共享规则库全量:
node ~/.dsh/rules/check-rules.mjs与node ~/.dsh/rules/audit-docs.mjs(DSH 经~/.dsh/rules目录联接访问,OpenCode 直接用~/.config/opencode/,同一份实体)
判级速查
| 级别 | 典型线索 | 处理 |
| 高置信(退出码 1) | 旧版、原实现、以前是、曾是、不再支持、已移除、previously、used to、no longer | 删除;或改写成「曾经的坑 → 现修法」;确需引用时行内加 docs-audit:allow |
| 复查候选(仅提示) | 曾经、以前、旧版本、已删除、原本、重命名、deprecated | 人工判断:复发性坑写成「怎么防」保留;纯历史叙述删除 |
| 允许形式 | 同一行同时含「曾经的坑/复发/防回归」与「→/现修法/怎么防/判定方法」 | 无需处理 |
| 行内豁免 | docs-audit:allow | 规则词汇表、迁移说明等确需提及旧形态的场合 |
细节:围栏代码块与行内代码不参与扫描;引号内的高置信词自动降级为候选;--json 输出 { mode, scanned, high, candidates, allowed, warnings }。
修复三步
- 高置信命中:优先删除墓碑句;该坑可能复发时改写为「曾经的坑 → 现修法」,把篇幅给「现在怎么做」。
- 复查候选:逐条判断对当前代码有无操作/避坑价值——有(复发性坑)就写「怎么防」,没有就删。
- 确需保留旧形态词汇(词表、迁移说明):行内加
docs-audit:allow,并在完成报告中说明理由。
git 钩子(全局安装)
- 本机
core.hooksPath指向配置库git-hooks/,每次git commit自动对暂存区文档跑audit-docs.mjs --staged;配置库内提交加跑check-rules.mjs全套。 - 安装:
git config --global core.hooksPath <规则库>/git-hooks(本机 DSH$env:USERPROFILE\.dsh\rules\git-hooks;OpenCode~/.config/opencode/git-hooks,同一份) - 卸载:
git config --global --unset core.hooksPath - 仓库级
core.hooksPath(如 husky)优先于全局;--no-verify为单次逃生口。 - 误报需要放行时,用行内
docs-audit:allow;全局卸载留给钩子故障等例外。
完成前门禁
- 高置信命中清零后,再声明完成/提交(候选允许存在,但要在完成报告里写「已逐条判断」)。
- 结果写进完成报告「实际」列:
文档审计:N 高置信→0,M 候选已判断。
红线兜底
- 高置信命中处理完再提交;
git commit --no-verify只留给钩子故障等例外场景,使用后须在回复中说明。
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

