agentleFS
Sign inStart free
ozturkogluozturkoglu/fable-skills/skills/setup-audit/SKILL.md

setup-audit

Skill0 starsChanged 4 months agoMITView on GitHub

Audit and improve the Claude Code setup itself (CLAUDE.md files, skills, settings, memory, referenced paths) using the "retiring co-founder" method: inventory with evidence, verify each finding, fix safe items immediately, explain in plain language, distill any reusable method into a skill, and record what changed. Trigger when the user asks to audit/improve/review the setup ("audit my setup", "improve my setup", "review my setup", "before I retire"), before a frontier model becomes unavailable, or after a batch of process failures suggests the rules themselves are stale.

2 other projects use a version of this.

Adopted by

Most recently taken up by

  • Other versions2 projects use a version of this.

    Word for word · 0

    None found.

    Edited · 0

    None found.

    Wrote their own version · 2

    1. moberghr/mtk-agent-toolkit★ 9same name
    2. sunnypatneedi/skills★ 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 repository3 other files it gives its agents.

The file

We split this skill into the rules it gives an agent

5 rules. About 0 are common, 4 are distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈4

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Treat the auditing model as a co-founder about to retire: before leaving, it inspects the whole operation, fixes what it can, writes down …

    line 16
  2. Distinctive:

    Missing routing/guardrails: no rule for a decision that recurs (model choice, effort, delegation), or a guardrail that exists only inside …

    line 44
  3. Distinctive:

    Every finding follows execution-guardrails rule 1: name the file+line that proves it.

    line 50
  4. Very specific:

    Ask: "Did this audit use a method worth reusing?

    line 64
  5. Distinctive:

    Update persistent memory (and the project's own running-summary log, if one exists) with: what was fixed, what was proposed-and-pending, …

    line 72

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “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

Setup Audit — the retiring co-founder method

Treat the auditing model as a co-founder about to retire: before leaving, it inspects the whole operation, fixes what it can, writes down WHY, and converts its know-how into artifacts the successors (cheaper models) can run without it.

The audit output is not a report — it is applied fixes + distilled skills. A finding that ends as prose helps nobody next session.

Stage 1 — Inventory (evidence, not memory)

Actually open, in this run:

  • ~/.claude/CLAUDE.md (global rules) and every project .claude/CLAUDE.md
  • ~/.claude/skills/*/SKILL.md — names, descriptions, cross-references
  • ~/.claude/settings.json + project settings.local.json (permissions, hooks)
  • memory index (MEMORY.md) and any memory file the audit touches
  • project knowledge entry points (e.g. a running-summary file, docs index)

Failable check: list what was actually read. Anything judged but not opened is marked "not inspected", never assumed fine.

Stage 2 — Findings (verify-before-flag)

Hunt specifically for:

  1. Stale facts: paths that no longer exist, machine/user names that changed, rules referencing removed systems, dates that silently expired.
  2. Broken references: a skill/doc/companion referenced somewhere but absent on disk; memory entries pointing at renamed files.
  3. Contradictions: two rules that cannot both be followed; global vs project rule conflicts.
  4. Missing routing/guardrails: no rule for a decision that recurs (model choice, effort, delegation), or a guardrail that exists only inside one mode instead of always-on.
  5. Dead weight: permissions/rules for workflows that no longer exist (flag; deleting is the user's call).

Every finding follows execution-guardrails rule 1: name the file+line that proves it. Classify each: FIX-NOW (safe, reversible) / PROPOSE (needs user decision) / DEFER (real but out of scope — note where).

Stage 3 — Apply and explain

  • Apply FIX-NOW items immediately; keep edits minimal and dated.
  • For each fix, one plain-language line: what was wrong → what it is now → why it matters. No jargon the user would have to decode.
  • PROPOSE items go in one batched list at the end (execution-guardrails rule 2), each with its evidence.

Stage 4 — Distill (the step that outlives the model)

Ask: "Did this audit use a method worth reusing? Did the user admire an output without knowing how it was made?" If yes, convert the method into a skill (~/.claude/skills/<name>/SKILL.md) or update the one that should have covered it. The test: could a mid-tier model rerun this audit next month from the skill alone, with this session gone?

Stage 5 — Record

Update persistent memory (and the project's own running-summary log, if one exists) with: what was fixed, what was proposed-and-pending, where new artifacts live. Dates absolute, never relative.

Failable check for the whole run: re-list the created/edited files with a fresh directory listing — every claimed artifact must appear in it.

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.

Check your setup, free