AugmentClaude

Skill Stocktake

Audit AGF's internal skills for content quality and relevance.

Installation

  1. Make sure Claude is on your device and in your terminal.

    Skills load from ~/.claude/skills/ when Claude Code starts up — so you need it on your machine first. If you don't have it yet, install it once with the command below, then run claude in any terminal to verify.

    One-time setup
    npm i -g @anthropic-ai/claude-code

    Already have it? Skip ahead.

  2. Paste into Claude Code or into your terminal.

    This copies the whole skill folder into ~/.claude/skills/agf-skill-stocktake-pcliangx/ — the SKILL.md plus any scripts, reference docs, or templates the skill ships with. Safe default: works for every skill.

    Faster alternative (instruction-only skills)

    Skips the clone and grabs only the SKILL.md file. Don't use this if the skill ships Python scripts, reference markdowns, or asset templates — they won't be downloaded and the skill will fail when it tries to load them.

    Quick install (SKILL.md only)
    Sign up to copy
  3. Restart Claude Code.

    Quit and reopen Claude Code (or any other agent that loads from ~/.claude/skills/). New skills are picked up on startup.

  4. Just ask Claude.

    Skills auto-activate when your request matches the skill's description — no slash command needed. Trigger phrases live in the skill's own frontmatter; you can read them in the “What this skill does” section above.

Prefer to read the source first? Open on GitHub.

When Claude uses it

Use when PL / tech-lead audits the content quality of AGF's own skills (`.claude/skills/agf-*/SKILL.md`). Provides the 5-dimension checklist (Actionability / Scope fit / Uniqueness vs CLAUDE.md+standards / Currency of pointers / Gate-binding validity), 5-verdict rubric (Keep/Improve/Update/Retire/Merge), reason-quality bar (反敷衍), and output skeleton. Triggered each MINOR release retro or manually.

What this skill does

Skill Stocktake(AGF 自有 skill 内容质量审计)

Use this skill when:

  • MINOR release retro 时审 AGF 自有 skill 是否腐化
  • 手动盘点 skill 质量(感觉 skill 集膨胀 / 指针过期 / 内容重叠)
  • 新增 / 删除 skill 后回归一次确保无重叠

What it is — and is not

Skill Stocktake 审计 AGF 自有治理型 skill 的内容质量——不是语法(lint-all.sh 管)、不是 SSOT drift(gen-roles.py --check 管)。它审 skill 是否仍 actionable / scope-fit / unique / 指针新鲜 / 绑定有效 gate。

不审:第三方 skill(docx / pptx,Anthropic 提供,非 AGF 治理资产)。

Scope

.claude/skills/agf-*/SKILL.mdls .claude/skills/agf-*/SKILL.md 实际为准,不在文档硬编码计数(守 verified-facts.md 纪律)。

5 维 checklist(每 skill 逐条核)

AGF 在 ECC 4 维上加 1 维 Gate 绑定(AGF 治理型 skill 的核心——每个 agf-* 都该绑一个受治理流程 + gate):

  1. Actionability:有可立即执行的步骤 / 命令 / 代码示例(不是空泛原则)
  2. Scope fit:name / trigger(description)/ 内容对齐,不过宽不过窄
  3. Uniqueness:内容不被 CLAUDE.md / standards/ / 另一 skill 替代(守 SSOT,不重复)
  4. Currency(指针新鲜度):引用的 CLAUDE.md / standards/*.md / docs/adr/*.md 路径仍存在 + 内容同步(用 rg / [ -f ] 验路径,抽查引用内容是否漂移)
  5. Gate 绑定有效性(AGF 特有):skill 是否仍绑一个有效 gate / SSOT 命令(如 agf-deploying-uat 绑部署门 + /agf-deploy-uatagf-writing-change 绑变更文件夹 + agf-spec-validate)。无 gate 的 skill 要么是知识型(说明)要么是孤儿(候选 Retire)

5 档 verdict

Verdict含义
Keep5 维全过,仍有效
Improve值得保留但需具体改进(指明哪段 / 什么改 / 目标规模)
Update引用的技术 / 路径过期(rg / WebSearch 验证后更新)
Retire低质量 / 过时 / cost-asymmetric(说明什么覆盖了它)
Merge into [X]与另一 skill 重叠(指明目标 + 整合什么内容)

Reason 质量要求(反敷衍)

评估是整体 AI 判断(非数值 rubric)。但 reason 字段必须自包含、可决策:

  • ❌ 不写 "unchanged" / "overlaps with X" 这种空话
  • Retire:(1) 发现什么具体缺陷 (2) 什么覆盖了同一需求
    • Bad: "Superseded"
    • Good: "agf-writing-prd 已 v6.9.0 弃用,agf-writing-change 覆盖全部场景 + delta + archive;无独有内容残留"
  • Merge:目标 skill + 整合什么内容
  • Improve:哪段、什么改、目标规模
  • Keep(仅 mtime 改):重述原 verdict 依据,不写 "unchanged"

流程

1. Inventory

ls .claude/skills/agf-*/SKILL.md

2. 逐 skill 评估

对每个 skill:Read SKILL.md → 5 维核对 → 填 verdict + reason。

Currency 维必须验路径(静默腐化高发点):

# 验 skill 引用的路径存在
rg -oE '\[.*?\]\(([^)]+)\)' .claude/skills/agf-<name>/SKILL.md | ...   # 抽链接
[ -f <引用路径> ] && echo OK || echo STALE

skill 数少(~15),一批跑完,不需要 chunk / resume(ECC 的 chunk 是为 261 skill 设计,AGF 裁剪掉)。

3. 输出

docs/reviews/skill-stocktake-<YYYY-MM-DD>.md

  • frontmatter(reviewer / date / scope: agf-* / skill_count
  • 汇总表(Skill | Verdict | Reason 一行)
  • 按 verdict 分组详情
  • 改进项(Improve / Update / Retire / Merge 的具体 action 清单)

4. PL 决策

Retire / Merge 涉及删 / 合 skill,PL 确认后才动(不自动删——守 agf-writing-change 变更门,删 skill 是 REMOVED Requirement 级变更)。Improve / Update 由 PL 派 maintainer 或 tech-lead 执行。

触发节奏

  • 每 MINOR release retro:作为 retro 的可选步骤(见 agf-running-release-retro
  • 手动:PL / tech-lead 觉得 skill 集膨胀 / 指针过期时随时跑
  • PATCH 不跑:skill 腐化慢,MINOR 周期够
  • MAJOR release 前:强烈建议跑一次(大版本是清理窗口)

Anti-patterns

  • ❌ 把 stocktake 当 lint 跑(语法 / drift 是 lint-all.sh + gen-roles.py --check 的事)
  • ❌ 审第三方 skill(docx / pptx 非 AGF 治理资产)
  • ❌ Retire / Merge 不经 PL 确认直接删 / 合(违反变更门)
  • ❌ reason 写 "unchanged" / "overlaps"(反敷衍要求)
  • ❌ 跳过 Currency 维的 rg 验证(指针过期是静默腐化,不验 = 没审)
  • ❌ 硬编码 skill 计数(以 ls 实际为准,守 verified-facts.md

Related skills