agentleFS
Sign inStart free
gonimargonimar/claude-web-studio/CLAUDE.md

claude-web-studio

CLAUDE.md4 starsChanged 45 days agoMITView on GitHub

- The repository root is the plugin root: agents/, skills/, hooks/, docs/, rules/, templates/, testing/, tests/, evals/; .claude-plugin/ holds plugin.json and marketplace.json (this repository is its own marketplace). Commit gate: bash tests/run-all.sh (syntax, structure linter, hook smoke tests, installer tests, then claude plugin validate . and claude plugin validate .claude-plugin/plugin.json — the first checks the marketplace manifest, the second the plugin; both must pass). - Path mapping: the kit's docs/ and rules/ become the project's .claude/docs/ and .claude/rules/ (install.sh, /init);…

The file

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

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

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈0

    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. Very specific:

    Every new skill/agent gets a spec in testing/ (from testing/templates/) and a catalog.yaml entry.

    line 23

≈ “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

claude-web-studio — rules for developing the kit itself

  • The repository root is the plugin root: agents/, skills/, hooks/, docs/, rules/, templates/, testing/, tests/, evals/; .claude-plugin/ holds plugin.json and marketplace.json (this repository is its own marketplace). Commit gate: bash tests/run-all.sh (syntax, structure linter, hook smoke tests, installer tests, then claude plugin validate . and claude plugin validate .claude-plugin/plugin.json — the first checks the marketplace manifest, the second the plugin; both must pass).
  • Path mapping: the kit's docs/ and rules/ become the project's .claude/docs/ and .claude/rules/ (install.sh, /init); skills, agents and rules are written against the project paths, so .claude/docs/stack-reference/go.md in a skill is docs/stack-reference/go.md here.
  • Three test trees: tests/ — the kit's own bash/python tests (run-all); testing/ — behavioural specs for skills and agents (/skill-test, catalog.yaml); evals/ — behavioural evals (claude plugin eval).
  • Local dev loop: claude --plugin-dir <path-to-this-checkout> run from a scratch project (install.sh --new /tmp/x --with-testing for copy mode; the marketplace route is in CONTRIBUTING.md § 1).
  • Hooks: no jq (the jget helper); a warning that must reach the model is JSON via warn, never permissionDecision; exit 2 blocks on PreToolUse only; every hook is registered in both hooks/hooks.json and templates/settings.json and gets a case in tests/hooks.sh.
  • Talk to the maintainer in whatever language they use; tracked files stay in English (README translations in docs/readme/). Nothing personal or project-specific goes into tracked files (no real hosts, names, home paths, private project names). Local notes and development history live in dev/ (gitignored).
  • Facts about stack versions belong only in docs/stack-reference/*.md with updated: and sources:; refresh with /stack-update run here.
  • Every new skill/agent gets a spec in testing/ (from testing/templates/) and a catalog.yaml entry; run /skill-test static all after editing skills or agents. Critical skills also have eval cases in evals/ (claude plugin eval, see testing/README.md § Evals); a changed rule gets a case. A new command is added to the command lists of README.md and every docs/readme/README.*.md (run-all fails with "commands not documented" otherwise).
  • The principles in templates/CLAUDE.md.template are cited by number from rules, skills and the playbook: extend the list, never renumber it; a cut principle stays as a one-line pointer with its number.
  • Release: bump version in .claude-plugin/plugin.json, add a CHANGELOG.md entry, commit chore(release): vX.Y.Z. The installer must never touch project data — check install.sh <scratch-repo> --dry-run after changing it. Extension guides: CONTRIBUTING.md.

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