agentsclimarketplace

Skills audit

Skill JochenYang/Jochen-ai-rules/skills/skills-audit

List all installed skills with line counts and analyze for improvement opportunities. Use when user wants to review, optimize, or clean up their skill collection.From its SKILL.md

Install
npx -y skills add JochenYang/Jochen-ai-rules --skill skills-audit

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

  • 20 stars20 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

SKILL.md

3.2 KB, as published. Nobody here has run it

Skills Audit Skill

Analyze all installed skills to find optimization opportunities.

Workflow

  1. Scan skills directories:
    • Global: ~/.claude/skills/
    • Project: .claude/skills/
  2. Collect metadata for each skill:
    • Name from SKILL.md frontmatter
    • Line count of SKILL.md
    • File count in skill directory
  3. Analyze for improvement opportunities:
    • Overlapping scopes: Similar functionality across skills
    • Token efficiency: Skills with excessive line counts
    • Consolidation candidates: Skills that could be merged
  4. Present findings in structured format
  5. Ask user: Which skills to review or optimize?

Analysis Criteria

Overlapping Scopes

  • Two or more skills handling similar domains
  • Redundant triggers or descriptions
  • Conflicting instructions

Token Efficiency

  • SKILL.md > 500 lines: Consider splitting
  • Excessive examples or explanations
  • Content that could move to references/

Consolidation Candidates

  • Skills frequently used together
  • Related skills in same domain
  • Small skills (<50 lines) that could merge

Output Format

# Skills Audit Report

## Summary
- Total skills: X
- Total lines: Y
- Project skills: A
- Global skills: B

## Skills List

| Skill | Location | Lines | Files |
|-------|----------|-------|-------|
| name | global/project | N | M |

## Findings

### Overlapping Scopes
| Skill A | Skill B | Overlap |
|---------|---------|---------|
| a.md | b.md | description |

### Token Efficiency Issues
| Skill | Lines | Recommendation |
|-------|-------|---------------|
| name | 600+ | Split into references/ |

### Consolidation Candidates
| Skills | Rationale |
|--------|-----------|
| a + b | Frequently used together |

## Action Items
[ ] Review skill: name
[ ] Split skill: name
[ ] Merge skills: a + b

Present findings to user and ask which to implement

Boundaries

  • Focus on skill inventory quality, overlap analysis, and optimization guidance.
  • Do not merge, split, or rewrite skills automatically without owner confirmation.
  • Treat token-efficiency suggestions as tradeoffs, not mandatory reductions.

When NOT to Use

  • Implementing code features unrelated to skill inventory analysis
  • Running destructive skill cleanup actions without explicit owner decision
  • Replacing hands-on quality/security review with inventory-only metrics

Escalation Rules

Pause and ask the owner before:

  • recommending deletion of heavily used skills
  • consolidating skills with partially overlapping but distinct responsibilities
  • changing invocation patterns that could break existing workflows

Final Output Contract (MANDATORY)

Every use of this skill should end with:

  1. Skill Fit - why a skills audit is appropriate now
  2. Primary Deliverable - inventory summary and optimization recommendations
  3. Execution Evidence - directories scanned, counts gathered, and overlaps found
  4. Risks / Open Questions - migration cost, ambiguity, or retained complexity
  5. Next Action - the highest-value cleanup or review step

Keep looking

Skills are one crate of 627,801. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.