uinaf/skill-audit
Audit existing skills with Tessl scoring, metadata and trigger-coverage checks, repo conventions, and skill-authoring best practices. Use when creating or revising a skill, triaging weak self-activation, or comparing a skill against source-repo guidance such as `AGENTS.md`, `CLAUDE.md`, or repo rules, plus external skill guidance. Do not use to verify general application code or to rewrite unrelated docs.
97
98%
Does it follow best practices?
Impact
97%
1.05xAverage score across 3 eval scenarios
Advisory
Suggest reviewing before use
Quality
Discovery
100%Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.
This is a strong, well-crafted description that clearly defines a specific niche (skill auditing), lists concrete actions, provides explicit 'Use when' and 'Do not use' clauses, and includes natural trigger terms. The negative boundary ('Do not use to verify general application code') is a particularly effective touch for disambiguation. Minor improvement could include mentioning file extensions like .md for skill files, but overall this is excellent.
| Dimension | Reasoning | Score |
|---|---|---|
Specificity | Lists multiple specific concrete actions: 'Tessl scoring', 'metadata and trigger-coverage checks', 'repo conventions', and 'skill-authoring best practices'. These are concrete, identifiable audit activities rather than vague abstractions. | 3 / 3 |
Completeness | Clearly answers both 'what' (audit skills with scoring, metadata checks, conventions, best practices) and 'when' (creating/revising skills, triaging weak self-activation, comparing against repo guidance). Also includes explicit 'Do not use' guidance, which strengthens the 'when' clause. | 3 / 3 |
Trigger Term Quality | Includes strong natural trigger terms users would say: 'creating or revising a skill', 'weak self-activation', 'AGENTS.md', 'CLAUDE.md', 'repo rules', 'skill guidance'. Also includes negative triggers ('Do not use to verify general application code') which help with disambiguation. | 3 / 3 |
Distinctiveness Conflict Risk | Highly distinctive niche — auditing skills specifically with Tessl scoring and skill-authoring conventions. The explicit 'Do not use' clause further reduces conflict risk with general code review or documentation skills. Unlikely to be confused with other skills. | 3 / 3 |
Total | 12 / 12 Passed |
Implementation
85%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
This is a well-structured audit skill that follows its own advice about progressive disclosure and lean SKILL.md content. The workflow is clearly sequenced with validation checkpoints (Tessl rerun) and a defined output format. The main weakness is that actionability could be stronger with more concrete, executable examples showing expected Tessl output or a complete audit walkthrough.
Suggestions
Add a concrete example of a complete audit footer output so Claude knows exactly what the expected deliverable looks like.
Include a brief executable example showing the full Tessl command and a sample JSON output snippet to make the 'Run Tessl first' step more copy-paste actionable.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The skill is lean and efficient throughout. It assumes Claude's competence, avoids explaining what Tessl is beyond a one-liner, and uses terse bullet points and labeled sections. Every section earns its place with no padding or redundant explanation. | 3 / 3 |
Actionability | The skill provides concrete CLI commands (e.g., `npx tessl skill review --json skills/<name>`) and specific examples (weak vs stronger naming), but lacks fully executable end-to-end examples. The workflow steps are described at a guidance level rather than with copy-paste-ready command sequences or concrete output examples. | 2 / 3 |
Workflow Clarity | The six-step workflow is clearly sequenced with explicit ordering (run Tessl first, then audit discovery, shape, disclosure, repo fit, synthesize). It includes validation via Tessl rerun after edits, concrete failure signs to check for, and a defined 'done' state (compact audit footer with specific labeled lines). | 3 / 3 |
Progressive Disclosure | The skill keeps the core workflow in SKILL.md and explicitly delegates dense content to `references/scorecard.md` and `references/best-practices.md`. References are one level deep, clearly signaled with relative links, and the skill even instructs when to move content into references vs scripts. | 3 / 3 |
Total | 11 / 12 Passed |
Validation
100%Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.
Validation — 11 / 11 Passed
Validation for skill structure
No warnings or errors.
Reviewed
Table of Contents