catalog
Skill2 starsChanged 56 days agoMITView on GitHub
Use when starting work in an unfamiliar repo, setting up hygiene skills, or unsure which skills a task needs. Recommends and installs a subset.
5 other projects use a version of this.
Adopted by
Most recently taken up by
- weininghui/skillstheir own version, first seen Oct 30, 2026
- summationai/summation-plugintheir own version, first seen Oct 20, 2026
- Pearl1531/Archetype-talktheir own version, first seen Oct 13, 2026
- LiberatedPixelCup/Universal-LPC-Spritesheet-Character-Generatortheir own version, first seen Oct 8, 2026
- ksprashu/agent-skill-forgetheir own version, first seen Sep 10, 2026
Other versions5 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 5
- LiberatedPixelCup/Universal-LPC-Spritesheet-Character-Generator★ 1.9ksame name
- ksprashu/agent-skill-forge★ 3same name
- Pearl1531/Archetype-talk★ 2same name
- weininghui/skills★ 1same name
- summationai/summation-plugin★ 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.
In this repository20 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- artifact-handoffSkill
- changelog-qualitySkill
- change-splittingSkill
- code-structureSkill
- comment-qualitySkill
- evidence-disciplineSkill
- factual-integritySkill
- language-style-samplingSkill
- llm-tellsSkill
- plan-iteration-gateSkill
- pre-commit-checkSkill
- project-discoverySkill
- repo-hygieneSkill
- self-review-gateSkill
- skill-authoringSkill
- socratic-specSkill
- systematic-debuggingSkill
- upstream-hygieneSkill
The file
We split this skill into the rules it gives an agent
4 rules. 0 are common, 3 are distinctive, and 1 is very specific to this project.
0
Common
Most of the 5 edited versions we read keep these.
3
Distinctive
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
- Distinctive: line 16
Use when user says "set up hygiene", "make my AI write better PRs", "what skills exist?", or wants onboarding.
- Distinctive: line 26
Ask 2-3 discovery Qs (clarify choices array, never embed options inside question text):
- Very specific: line 35
Always unless personal repo only: upstream-hygiene + factual-integrity CLI/library: + comment-quality + code-structure + changelog-quality …
- Distinctive: line 51
No auto-install ever — confirmation gate.
“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
Catalog — What's Available + Install
Use when user says "set up hygiene", "make my AI write better PRs", "what skills exist?", or wants onboarding.
What this skill does
- Load inventory:
skills/dirs — read frontmatter description + first para — positive framing, what skill provides not what isn't.references/index.md— cliff notes on external repos/tools (10 entries v0.2).references/generic-principles.mdsummary.AGENTS.mdcanonical +CLAUDE.mdshim note per ADR 0001.
Ask 2-3 discovery Qs (clarify choices array, never embed options inside question text):
- What type of project: library, CLI tool, web service, docs, kernel/module?
- Where does code ship: personal repo only, OSS fork/PR, company monorepo, mail list patch series?
- Subject spec: does project already have conventional commits / commitlint / own subject style per CONTRIBUTING? — decides between
conventional-commitsreference vs changelog-quality GC-10. - Agent home where you want install:
~/.claude/skills,~/.hermes/skills,~/.agents/skills(Claw),~/.openclaw/skills,~/.config/opencode/skills,AGENTS.mdgeneric,.cursor/rules,.github/copilot-instructions.md?
Recommend 5-7 core + 2-3 references based on answers:
Always unless personal repo only:
upstream-hygiene+factual-integrityCLI/library: +comment-quality+code-structure+changelog-quality+change-splittingAmbiguous task incoming: +socratic-spec+plan-iteration-gate+systematic-debuggingPRs to OSS: +project-discovery+pre-commit-checkPre-commit self-check:self-review-gate4Q (a) wrong (b) better way (c) missing (d) boundaries +llm-tellsNew language/project with sparse style: +language-style-sampling+ refgoogle-eng-practicesConventional commits detected: refconventional-commits+lint-chainStacked PR workflow: refstacked-pr+ corechange-splitting- Present as checklist with one-sentence TL;DR per skill (cite source file path, not memory) + positive capability. Then offer install.
Install gate:
- List:
scripts/install.sh --list— probes agent homes. - Install:
scripts/install.sh --install <comma-list>— requires explicit yes. - Generic agent: append
cat skills/<name>/SKILL.mdsnippet + reference link intoAGENTS.mdwith provenance comment<!-- from code-hygiene skills/<name> -->. - No auto-install ever — confirmation gate.
- List:
Output format
## Available — core (shipped, 15 skills v0.2)
- upstream-hygiene — public-only commits/comments class-based ban placeholder-only
- factual-integrity — never invent...
- change-splitting — one logical change per commit, 200-line seam trigger...
- systematic-debugging — 4-phase root cause...
...
## Available — references (cliff notes, not vendored, 10 entries)
- google-eng-practices — small CLs, WHY, reviewer quality
- conventional-commits — type(scope): subject spec + commitlint
- stacked-pr — ghstack/jj/git-branchless
...
## Recommended for your project (5-7 core + 2-3 refs)
- [ ] upstream-hygiene — ...
- [ ] ...
Invoke: ./scripts/install.sh --install <list>Verification — 4Q
(a) Wrong? Does catalog list skill that doesn't exist on disk, invents description, or violates upstream-hygiene placeholder-only?
(b) Better way? Existing generic mechanism — project already has AGENTS.md with rules — reuse and extend not duplicate? State if none beats with tradeoff.
(c) Missing? Rollback plan if install overwrites existing AGENTS.md, observability per install step (file exists + content hash), failure notification if skill dir not writable, what unblocks downstream (PR creation), what stays out of scope (full language lint config).
(d) Boundaries explicit? Which skills core vs reference cliff notes not vendored, dep version pinned, which checks org-specific private overlay ~/.config/hygiene/extra-check.sh, scope positive: "provides installable skills + probe + confirmation, not auto-install".
Anti-patterns
- Listing everything without recommendation — choice paralysis defeats purpose.
- Auto-installing without asking — never.
- Defining by negation "kernel-free" in description — positive capability per positive framing rule.
- Installing references as vendored code — only cliff notes + link, user opts in.
Related
project-discoveryStep0 finds upstream CONTRIBUTING that bounds recommendation.upstream-hygieneowns banned classes — verification gate uses its check-hygiene.sh.self-review-gate4Q own recommendation before presenting.
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.