Documents

Skills Audit

Try it

Audit all installed skills to identify duplicates, platform-mismatched skills, and maintenance candidates. Use when: (1) the user asks to "clean up skills",...

What it does

Audit all installed skills to identify duplicates, platform-mismatched skills, and maintenance candidates. Use when: (1) the user asks to "clean up skills", "check for duplicate skills", "audit installed skills", or "find skills to remove"; (2) performing periodic skill maintenance; (3) investigating skill clutter or skill conflicts. Scans all SKILL.md files in ~/.workbuddy/skills/, compares against a golden list of approved skills, classifies each as approved / unknown / platform-specific, and generates a structured audit report.

The skill document

Skills Audit

Audit all installed WorkBuddy skills, classify them, and generate a cleanup report.

Workflow

1. Run the audit script

python scripts/audit_skills.py

This scans ~/.workbuddy/skills/, reads each SKILL.md, and generates a skills-audit-YYYY-MM.md report.

Options:

  • --dir — custom skills directory (default: ~/.workbuddy/skills/)
  • --golden — custom golden list (default: references/golden-list.md)
  • --output — custom report output path

2. Interpret the report

The report classifies each skill into one of three categories:

StatusMeaningAction
[OK] ApprovedMatches the golden listKeep
[!] ReviewNot in golden list, not platform-specificManually review
[X] Platform:Designed for Clawdbot/Moltbot/OpenClawCandidate for removal

3. Present findings to user

Use present_files to show the report. For platform-specific or unknown skills, flag them and ask the user for confirmation before removing.

4. Remove redundant skills (after user confirmation)

rm -rf ~/.workbuddy/skills/

Classification Rules

  1. Golden list match: If the skill name or directory matches an entry in references/golden-list.md, it is Approved.
  2. Platform detection: SKILL.md is scanned for keywords (clawdbot, moltbot, openclaw, clawhub, CodeConductor.ai). If found and not in golden list, it is flagged as platform-specific.
  3. Otherwise: Classified as Unknown — requires manual review.

Bundled Resources

  • scripts/audit_skills.py — Python script that performs the scan and generates the report
  • references/golden-list.md — Canonical list of approved skills (edit this to add/remove approved skills)

Related skills

Audit a target SKILL.md against the Agent Skills specification and generate a Chinese HTML report. Use when the user asks to check, audit, review, or optimiz...

21 installs

Runtime-neutral skill operations suite. Input a skill directory, slug, marketplace goal, or release plan; output quality checks, safety boundaries, naming/su...

4 installs

Agent skill quality checker. Input a skill directory or skill files; output trigger clarity, metadata issues, examples, safety boundaries, installability, po...

36 installs

Audit another Codex skill for structural compliance, trigger quality, instruction clarity, reuse of scripts or references, and overall maintainability. Use w...

16 installs

对已存在 Skill 做 8 维度全面体检(结构/安全/触发/有效性/竞争/平台/文档/代码质量)。说 技能审计/审计技能/技能体检 时触发。支持成熟度分级+4确认点+整改+回归审计。绝不自动发布。Do NOT use for creating skills or publishing to platforms.

Agent skill recommender. Input a user need, task description, or existing skill list; output best matching skills, install rationale, duplicate/merge candida...

37 installs