repo-hygiene
Skill252 starsChanged 2 months agoMITView on GitHub
Audit and clean up the repository — remove dead files, fix naming inconsistencies, update .gitignore, and ensure the repo is well-organized for OSS contributors. Does NOT modify source code logic.
15 other projects use a version of this.
Adopted by
Most recently taken up by
- Smaui94/ClaudeSkillCollectiontheir own version, first seen Nov 8, 2026
- robinbril/claude-code-harnesstheir own version, first seen Nov 1, 2026
- yunsii/wezdecktheir own version, first seen Oct 25, 2026
- arndvs/ctrlshfttheir own version, first seen Oct 15, 2026
- talas9/anti-halltheir own version, first seen Oct 13, 2026
- Donemeb/Donemeb-Skillstheir own version, first seen Oct 13, 2026
Used with
Projects with it also use systematic-debugging.
Other files people use alongside this one.
Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
OpenAISkillsystematic-debugging
Use when encountering any bug, test failure, or unexpected behavior, before proposing fixes
Also in 3 projects that use this
Other versions15 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 15
- marcfargas/skills★ 4same name
- marcoguillermaz/Tierward★ 4same name
- yunsii/wezdeck★ 4same name
- robinbril/claude-code-harness★ 3same name
- rikvanriel/code-hygiene★ 2same name
- talas9/anti-hall★ 2same name
- Asher-Plihal/claude-skills★ 1same name
- jedarden/jeds-curated-skills★ 1same name
- SamuelNDCE/ai-workbench★ 1same name
- SamuelNDCE/claude-super-skill-library★ 1same name
5 more not listed
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 repository7 other files it gives its agents.
- AGENTS.mdAGENTS.md
- ci-fix-and-verifySkill
- full-code-reviewSkill
- perf-loopSkill
- review-responseSkill
- upgrade-svelteSkill
- verify-svelte-compatSkill
The file
We split this skill into the rules it gives an agent
2 rules. 0 are common, 0 are distinctive, and 2 are very specific to this project.
0
Common
Most of the 14 edited versions we read keep these.
0
Distinctive
This author's own choices. Worth a look.
2
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 65
No root [[bin]]; dev binaries are crates/rsvelte_devtools/src/bin/*.rs and crates/rsvelte_core/src/bin/.
- Very specific: line 108
Run sections 1–10 in order.
“Common” is measured: kept by at least 6 in 10 of the 14 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
Repository Hygiene
Audit repository structure, naming, configuration and documentation. Never modify source code logic. Deletions need user confirmation.
Never delete
| What | Why |
compatibility/*known-failures*.json, *-oracle-excluded.json, *-mechanisms.json, *.provenance.json, attribution-pending.json, KNOWN-FAILURES.md, GATES.md | shrink-only CI ratchets and their justifications |
compatibility/{pattern-corpus,lint-adversarial,lint-env,check-fixtures,lsp-fixtures}/ | checked-in gate inputs |
submodules/* (every entry in .gitmodules) | pinned corpus sources and upstream references |
fixtures/, target*/, pkg*/, compatibility/{sources,expected,actual,fmt,…}/, .corpus-cache/ | generated and gitignored; reclaim only with pnpm run corpus:clean, never git rm |
pgo/rsvelte.profdata, upstream_issues/ | shipped-build input; index gated by scripts/ci/check-upstream-issues.mjs |
Checklist
1. Dead and generated files
find . \( -name "*.backup" -o -name "*.bak" -o -name "*.orig" -o -name "*~" -o -name ".DS_Store" -o -name "Thumbs.db" \) \
-not -path "*/node_modules/*" -not -path "./.git/*" -not -path "*/target/*" -not -path "./submodules/*"
ls -la benchmark-results.json .benchmark-files.txt npm 2>/dev/null # should be gitignored / absent
find . -type d -empty -not -path "*/node_modules/*" -not -path "./.git/*" -not -path "*/target/*" -not -path "./submodules/*"Dead file → git rm. Generated but tracked → add to .gitignore, then git rm --cached.
2. .gitignore coverage
Confirm each is present; add what is missing:
| Group | Patterns |
| OS / editor | .DS_Store, Thumbs.db, *.swp, *.swo, *~ |
| Rust | /target, **/*.rs.bk |
| Generated | benchmark-results.json, .benchmark-files.txt, /fixtures/, /pkg, /npm/ |
| Deps | node_modules/, .pnpm-store/ |
| IDE / backup | .idea/, .vscode/, *.code-workspace, *.backup, *.bak, *.orig |
3. Unused scripts
Scripts live in nested dirs (scripts/{bench,ci,compat-corpus,compat-lsp,dev,diff,fixtures,perf,release,reports}/).
for f in $(find scripts -type f \( -name '*.mjs' -o -name '*.sh' \) -not -path '*/node_modules/*'); do
name=$(basename "$f")
hits=$(grep -rl --exclude-dir=node_modules --exclude-dir=target "$name" package.json README.md AGENTS.md .github scripts crates 2>/dev/null | grep -v "^$f$" | wc -l)
[ "$hits" -eq 0 ] && echo "UNREFERENCED: $f"
doneReferenced anywhere → keep. Debugging utility (e.g. scripts/diff/compare-parsers.mjs) → keep and
document. Unreferenced and undocumented → [ask].
4. Unused binaries
No root [[bin]]; dev binaries are crates/rsvelte_devtools/src/bin/*.rs and
crates/rsvelte_core/src/bin/. pnpm run report-orphan-test-binaries lists test targets
nothing references. A dev tool (profiler, benchmark_runner, …) is fine outside CI if documented.
5. Naming
- Scripts: kebab-case,
*.sh/*.mjs. - Rust files and dirs:
snake_case—find crates -name '*.rs' | grep -E '[A-Z]|-'must be empty. - Tests:
crates/*/tests/*.rsinsnake_case. - Non-Rust dirs:
kebab-case.
6. Documentation
CLAUDE.mdmust remain a symlink toAGENTS.md(ls -la CLAUDE.md);README.mdis for humans.- Stale references:
grep -oE '`[^`]+\.(rs|mjs|sh|json)`' README.md | tr -d '`' | while read f; do [ -e "$f" ] || echo "MISSING: $f"; done - OSS scaffolding:
CONTRIBUTING.md(present),.github/ISSUE_TEMPLATE/,.github/PULL_REQUEST_TEMPLATE.md. Report missing ones; do not create unless asked.
7. CI/CD
ls .github/workflows/; per workflow check referenced scripts exist and actions are pinned
(.pinact.yaml is the pin config). Run pnpm run check:workflow-triggers.
8. Dependencies
- Cargo:
cargo clippy --all-targets --all-features 2>&1 | grep unused; a dependency present in aCargo.tomlbut never imported →[ask]. package.json: everydevDependencyreferenced by some script, workflow or workspace package.
9. Build artifacts in git
git rev-list --objects --all | git cat-file --batch-check='%(objecttype) %(objectname) %(objectsize) %(rest)' \
| awk '/^blob/ {print $3, $4}' | sort -rn | head -20.wasm, .node and platform binaries must not be tracked.
10. Root cleanliness
ls -a vs git ls-tree --name-only HEAD: every root entry must be tracked, gitignored, or a
known local dir (target*/, node_modules/, fixtures/). Anything else (e.g. .measure/) → [ask].
Workflow
- Run sections 1–10 in order; report each item as
[ok],[fix](auto-fixable) or[ask](needs a decision). - Summarize findings in one table.
- Apply
[fix]items; ask the user about[ask]items and apply what is approved. - Commit:
git commit -m "chore: repository cleanup".
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.

