agentleFS
maks11zagermaks11zager/audit-claude/SKILL.md

audit-claude

Skill0 starsChanged 3 months agoMITView on GitHub

Ponovljiv audit Claude Code setupa — subagenti rudarijo pretekle seje za signale (ponavljanja, korekcije, trenja), križanje z inventarjem (skilli, agenti, hooki, memory, CLAUDE.md), rangirano poročilo z dokazi. Pokliči z /audit-claude vsake ~3 mesece ali ko imaš občutek, da workflow škripa. SAMO DIAGNOZA — nič ne implementira brez potrditve.

1 other project uses a version of this.

Adopted by

Most recently taken up by

  • Other versions1 project uses a version of this.

    Word for word · 0

    None found.

    Edited · 0

    None found.

    Wrote their own version · 1

    1. VladTheProgrammer/claude-asimov★ 0same 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.

The file

We split this skill into the rules it gives an agent

1 rule. About 0 are common, 1 is distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈1

    Distinctive · estimate

    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:

    Read-only razen poročila. Vsak predlog MORA imeti dokaz (seja + citat) — brez dokaza ne gre v poročilo.

    line 34

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 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

/audit-claude — audit Claude Code setupa

Metodologija dokazana 2026-07-06 (Fable 5). Deluje tudi z Opusom (rudarji: Sonnet). Prejšnje poročilo: ~/OneDrive/Desktop/claude-audit-2026-07.md — preberi ga NAJPREJ in primerjaj (kaj je bilo popravljeno, kaj se je vrnilo, kaj je novo). Glej tudi memory project_claude_code_audit_jul2026.

Faza 1 — Inventar (sam, brez subagentov)

  • Skilli: ls ~/.claude/skills/ (+ _archived), agenti: ~/.claude/agents/, commandi: ~/.claude/commands/
  • ~/.claude/settings.json: hooki, št. allow pravil, ali je v permission stringih kak živ API ključ
  • Testiraj vsak hook v živo (echo sample JSON | bash hook) — hook, ki molči, je lahko mrtev, ne zdrav! (Lekcija: confirm-git je bil 6+ mesecev tiho mrtev zaradi python stuba.)
  • Memory: ls memory/*.md | wc -l vs grep -c '^- \[' MEMORY.md → neindeksirane datoteke
  • CLAUDE.md veriga: izmeri skupno velikost importov (wc -c); >5k tokenov globalno = kandidat za de-import
  • Preveri directory-scoped skille v projektnih mapah (<projekt>/.claude/skills/) — se sploh naložijo tam, kjer delaš?

Faza 2 — Rudarjenje sej (agent session-miner, način A — signali)

  1. Seznam sej: find ~/.claude/projects -name "*.jsonl" -size +100k -printf "%T@ %s %p\n" | sort -rn | head -60 — interaktivne. Preskoči trenutno sejo. Majhne (<30KB) enotne seje = headless SDK klici → 1 subagent samo karakterizira vzorec.
  2. Razdeli ~40 sej med 5–6 paralelnih session-miner agentov (po datumih), vsakemu v prompt: seznam datotek + "način A — audit signali". Metodologija (digest skripta, tipi signalov, format) je v agentovi definiciji — ne podvajaj je v promptu.

Faza 3 — Clustering in odločitev

Združi signale ČEZ seje. Za vsak grozd: NOV SKILL / AVTOMATIZACIJA / POPRAVEK obstoječega / ČIŠČENJE / NIČ (enkratno). Križaj z inventarjem: signal brez pokritja → novo; konfiguracija brez signalov → čiščenje; konflikt pravil → popravek.

Faza 4 — Samokritika

Za 3 najšibkejše kandidate: "Ali dokazi to res podpirajo? Bi uporabnik to uporabljal vsak teden?" Šibke črtaj. Ne predlagaj novih skillov, če problem rešuje popravek obstoječega — sprawl je bil najdba audita #1.

Faza 5 — Poročilo

Piši v ~/OneDrive/Desktop/claude-audit-YYYY-MM.md, slovensko: Povzetek (5 alinej) → Batch A/B/C (rangirano po prihranek×pogostost/strošek) → za vsak kandidat: Problem / Dokaz (seja+citat!) / Predlog / Tip / Strošek-prihranek → sekcija "Kaj NE spreminjati" → primerjava s prejšnjim auditom (popravljeno/regresija/novo).

Pravila

  • Read-only razen poročila. Vsak predlog MORA imeti dokaz (seja + citat) — brez dokaza ne gre v poročilo. Prepovedani generični nasveti.
  • Po pregledu uporabnik reče "GO za X" → šele takrat implementiraj; po implementaciji posodobi status v poročilu + memory.

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.