agentleFS
paureispaureis/pau-skills/CLAUDE.md

pau-skills

CLAUDE.md2 starsChanged 21 days agoMITView on GitHub

This is a Claude Code plugin marketplace. Read this before changing anything. - skills/<plugin>/<skill>/SKILL.md: every skill, grouped by plugin. Bundled files (reference docs, scripts, templates) sit next to the SKILL.md that uses them. - hooks/<plugin>/: hook scripts and the hooks.json that registers them. hooks/lib/ is shared by all hooks. - .claude-plugin/marketplace.json: one entry per plugin plus the pau-skills bundle. Every entry is rooted at the repository ("source": "./", "strict": false). Never add .claude-plugin/plugin.json; it conflicts with that. - .claude/agents/: the…

The file

We split this CLAUDE.md into the rules it gives an agent

3 rules. About 0 are common, 1 is distinctive, and 2 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.

  • 2

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Follow docs/ADDING-A-SKILL.md for any new skill or hook.

    line 20
  2. Very specific:

    After changing a hooks.json or adding a plugin folder, run node scripts/build-marketplace.mjs.

    line 21
  3. Distinctive:

    Every skill works when installed on its own, without the plugin or any other skill.

    line 27

≈ “Common” is an estimate: at least 1 in 5 of the 26,980 CLAUDE.md 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
CLAUDE.md

Working on this repository

This is a Claude Code plugin marketplace. Read this before changing anything.

Layout

  • skills/<plugin>/<skill>/SKILL.md: every skill, grouped by plugin. Bundled files (reference docs, scripts, templates) sit next to the SKILL.md that uses them.
  • hooks/<plugin>/: hook scripts and the hooks.json that registers them. hooks/lib/ is shared by all hooks.
  • .claude-plugin/marketplace.json: one entry per plugin plus the pau-skills bundle. Every entry is rooted at the repository ("source": "./", "strict": false). Never add .claude-plugin/plugin.json; it conflicts with that.
  • .claude/agents/: the roster that geppetto installs, used by the batches. Change a roster agent in skills/agent-orchestration/geppetto/roster/ first, raise the roster number on every covenant marker, then update the installed copy. Never put agent files in a root agents/ folder: every plugin here would load it.
  • scripts/: repository tooling. tests/: node:test suites. docs/: install guide, origins, rules, how to add a skill.

Rules

  • Follow docs/ADDING-A-SKILL.md for any new skill or hook.
  • After changing a hooks.json or adding a plugin folder, run node scripts/build-marketplace.mjs. Never edit the skills or hooks fields in marketplace.json by hand.
  • Bundled files are referenced as ${CLAUDE_PLUGIN_ROOT}/skills/<plugin>/<skill>/<file> or ${CLAUDE_PLUGIN_ROOT}/hooks/<plugin>/<file>. A file in the skill's own folder can instead be ${CLAUDE_SKILL_DIR}/<file>, which also works when the skill is installed on its own, without the plugin. A test fails on any such path that does not exist.
  • Every skill works when installed on its own, without the plugin or any other skill. A mention of another skill is an optional extra that says what to do if it is not installed. A script a skill needs from another skill is copied into its folder. tests/standalone.test.mjs checks all three.
  • Scripts and hooks use the Node.js standard library only (Node 20+). A hook exits 0 on input it does not understand, unless failing closed is its purpose, and checks hookEnabled() from hooks/lib/config.mjs.
  • Skills are language-neutral: never assume JavaScript; cover the common ecosystems or detect the project's tooling.
  • Writing style: plain, direct English. No em or en dashes (a test enforces this), no marketing words.
  • Bump the version of every plugin whose files changed, and the pau-skills bundle, in marketplace.json, and add a line to CHANGELOG.md. When a hook changes, bump pau-skills-hooks as well.
  • The owner installs skills by uploading them to claude.ai (zips, one per skill, folder at the root) and hooks through pau-skills-hooks. After a merge that adds or changes a skill, hand over the zips for those skills, built on main with node scripts/pack-skills.mjs --since <main before the change>; see step 10 of "How to run a batch" in docs/backlog/how-to-run-a-batch.md. A new batch starts from the kickoff message in docs/backlog/kickoff.md.

Checks before every commit

npm test
npm run scrub
node scripts/build-marketplace.mjs --check
claude plugin validate . --strict

Prove a new test can fail with bash skills/verification/mutation-test/mutate.sh (commit first).

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.