audit-skills
Skill0 starsChanged 4 months agoMITView on GitHub
Audit Codex skills for DRYness, brevity, trigger quality, progressive disclosure, and unnecessary duplication of always-loaded or tool-provided guidance. Use when the user asks to review, trim, tighten, de-duplicate, or improve one or more skills, especially when deciding what belongs in SKILL.md versus nearby references, scripts, or assets.
9 other projects use a version of this.
Adopted by
Most recently taken up by
- ronniepinnell/caspertheir own version, first seen Nov 8, 2026
- MusserLab/lab-claude-skillstheir own version, first seen Oct 13, 2026
- tejastn10/workbenchtheir own version, first seen Oct 13, 2026
- RuoJi6/java-audit-skillstheir own version, first seen Oct 8, 2026
- pixee/pixee-clitheir own version, first seen Aug 18, 2026
- RuoJi6/audit-skillstheir own version, first seen Jun 16, 2026
Other versions9 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 9
- RuoJi6/audit-skills★ 1ksame name
- RuoJi6/java-audit-skills★ 1ksame name
- pixee/pixee-cli★ 31same name
- MusserLab/lab-claude-skills★ 6same name
- geemus/hax-yax★ 4same name
- Tharun-Balaji/audit-skills★ 4same name
- Yesterday-AI/skills★ 2same name
- ronniepinnell/casper★ 1same name
- tejastn10/workbench★ 1same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository5 other files it gives its agents.
- AGENTS.mdAGENTS.md
- audit-docsSkill
- audit-work-trackingSkill
- bootstrap-repo-standardsSkill
- ship-changeSkill
The file
We split this skill into the rules it gives an agent
12 rules. 0 are common, 9 are distinctive, and 3 are very specific to this project.
0
Common
Most of the 5 edited versions we read keep these.
9
Distinctive
This author's own choices. Worth a look.
3
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 20
Read only the targeted SKILL.md files and directly relevant nearby resources.
- Distinctive: line 24
Use only the sections that match the current audit.
- Distinctive: line 37
Prefer short bullets or fragments when clear.
- Very specific: line 40
Keep core workflow guidance in SKILL.md.
- Distinctive: line 44
Verify before calling something redundant.
- Distinctive: line 46
Do not assume a rule is always loaded.
- Distinctive: line 49
If the task includes fixes, make focused edits.
- Distinctive: line 50
Prefer deletion over rewriting, and rewriting over expansion.
- Very specific: line 51
If the skill is already concise and useful, say no change needed.
- Distinctive: line 63
Do not manufacture low-value findings.
- Distinctive: line 64
Do not preserve full sentences just because they read better.
- Distinctive: line 65
Do not create cosmetic edits to avoid an empty audit.
“Common” is measured: kept by at least 6 in 10 of the 5 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.
Files
Audit Skills
Primary lens: Review.
Goal
Review skills as compact operating guides, not mini handbooks.
Bias hard toward brevity. Keep only wording that changes behavior, routing, or risk.
Workflow
- Scope first.
- Identify the skill or skill set to review.
- Read only the targeted
SKILL.mdfiles and directly relevant nearby resources.
- Load the checklist after the scope is clear.
- Read references/audit-checklist.md.
- Use only the sections that match the current audit.
- Audit for trigger quality.
- Description says what the skill does and when to use it.
- Invocation is not broader than the workflow.
- Audit for DRYness and context discipline.
- Repeated always-loaded rules.
- Repo guidance that belongs in repo docs.
- Duplicated text across
SKILL.md,references/, and scripts.
- Audit for brevity first, not style polish.
- Cut filler, throat-clearing, duplicated framing, and obvious statements.
- Prefer short bullets or fragments when clear.
- Audit for progressive disclosure.
- Keep core workflow guidance in
SKILL.md. - Move details, examples, and checklists to nearby
references/. - Move repeatable deterministic work to
scripts/instead of spending tokens on long instructions.
- Keep core workflow guidance in
- Verify before calling something redundant.
- Point to where the duplicated guidance already lives.
- Do not assume a rule is always loaded.
- Edit or report.
- If the task includes fixes, make focused edits.
- Prefer deletion over rewriting, and rewriting over expansion.
- If the skill is already concise and useful, say
no change needed.
Output
- the skills reviewed
- findings ordered by impact
- file paths that support each finding
- whether each issue is trigger drift, repeated always-loaded guidance, repeated repo guidance, over-verbose body content, misplaced detail that should move to
references/orscripts/, duplication, or template residue - a final conclusion: fix now, defer with rationale, or no change needed
Do Not
- Do not manufacture low-value findings.
- Do not preserve full sentences just because they read better.
- Do not create cosmetic edits to avoid an empty audit.
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
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.

