agentleFS
gradigitgradigit/claude-skills/.claude/skills/auditing-skills/SKILL.md

auditing-skills

Skill0 starsChanged 9 months agoMITView on GitHub

Audits Agent Skills for spec compliance and best practices. Activates when user asks to "audit skill", "check skill", "validate skill", or mentions skill optimization. Do NOT use when creating, updating, or testing skills.

4 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

2 rules. 0 are common, 2 are distinctive, and 0 are very specific to this project.

  • 0

    Common

    Most of the 4 edited versions we read keep these.

  • 2

    Distinctive

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Check the skill's EVALUATIONS.md:

    line 94
  2. Distinctive:

    Update this skill when:

    line 171

“Common” is measured: kept by at least 6 in 10 of the 4 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Auditing Skills

Validates skills against Agent Skills specifications and best practices.

Workflow

- [ ] 1. Identify target skill (path or name)
- [ ] 2. Read SKILL.md and any referenced files
- [ ] 3. Run validation checks (see checklist.md)
- [ ] 4. Generate audit report
- [ ] 5. Offer to fix critical issues

Quick Audit

For a fast check, validate these critical items:

Frontmatter (Required)

FieldRule
nameLowercase, hyphens only, max 64 chars, matches directory name, no XML tags
descriptionNon-empty, max 1024 chars, third person, includes what + when, no XML tags

Frontmatter (Recommended)

FieldRule
licensePresent (MIT, Apache-2.0, etc.)
metadata.versionSemver string in quotes

Content

CheckRule
Line countSKILL.md body under 500 lines
WorkflowHas checklist of steps
ExampleHas concrete input/output
ReferencesOne level deep (no chains)

Audit Report Format

Generate report in this structure:

# Skill Audit: {skill-name}

**Path**: {path}
**Version**: {version or "not specified"}

## Summary

| Status | Count |
|--------|-------|
| Critical | X |
| Warnings | Y |
| Suggestions | Z |

**Overall**: PASS / NEEDS WORK / FAIL

## Critical Issues

{Issues that violate the spec - must fix}

## Warnings

{Best practice violations - should fix}

## Suggestions

{Optimizations - nice to have}

Evaluation Coverage

Check the skill's EVALUATIONS.md:

CheckSeverity
EVALUATIONS.md presentWarning (missing = NEEDS WORK)
3+ scenarios definedWarning
Given/When/Then formatWarning
At least one should-NOT-trigger scenarioWarning

For deep validation (running scenarios), use testing-skills.

Severity Levels

Critical (spec violations):

  • Missing required frontmatter fields
  • Invalid name format (uppercase, underscores, XML tags)
  • Name doesn't match directory
  • SKILL.md over 500 lines without progressive disclosure

Warning (best practice violations):

  • Missing license or version
  • No workflow checklist
  • No concrete examples
  • First/second person in description
  • Deeply nested references (>1 level)
  • Windows-style paths

Suggestion (optimizations):

  • Verbose explanations (Claude already knows X)
  • Changelog in SKILL.md instead of separate file
  • Reference files >100 lines without TOC
  • Missing self-evolution section

Common Issues

IssueFix
name: My-SkillUse lowercase: my-skill
name: my_skillUse hyphens: my-skill
Description: "I help you..."Use third person: "Helps users..."
Description: "Does stuff"Add specifics: what + when to trigger
SKILL.md 600 linesSplit into reference files
ref.md → detail.md → info.mdFlatten to one level

Example

Input: Audit ~/.claude/skills/my-skill

Output:

# Skill Audit: my-skill

**Path**: ~/.claude/skills/my-skill
**Version**: 1.0.0

## Summary

| Status | Count |
|--------|-------|
| Critical | 0 |
| Warnings | 2 |
| Suggestions | 1 |

**Overall**: NEEDS WORK

## Warnings

1. **No workflow checklist** - Add a workflow section with steps
2. **First person in description** - Change "I process files" to "Processes files"

## Suggestions

1. **Changelog in SKILL.md** - Move to separate CHANGELOG.md for token efficiency

Self-Evolution

Update this skill when:

  1. On spec change: Agent Skills spec updates → update checklist
  2. On new anti-pattern: Discover common issue → add to checks

Current version: 2.0.0. See CHANGELOG.md for history.

References

For complete validation criteria, see checklist.md.

Download now (ZIP)

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
What did you use it for? e.g. Reviewing pull requests in a Rails app

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.