agentleFS
Sign inStart free
gsscodergsscoder/claude-coding-agents/CLAUDE.md

claude-coding-agents

CLAUDE.md4 starsChanged 2 months agoMITView on GitHub

Definitions live in agents/<category>/ Narrow, domain-specific agents (single language/framework, e.g. dotnet-cs-expert) live in agents/<category>/specialists/ instead of flat under the category. Broad, general-purpose agents stay flat

The file

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

2 rules. About 0 are common, 2 are distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈2

    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:

    no memory; agents run stateless

    line 21
  2. Distinctive:

    no bold formatting; backticks allowed where appropriate

    line 22

≈ “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

AI guidance

Repo purpose

Collection of custom Claude Code agent and skill definitions. Typical session: create or update agent or skill definitions

Agent definitions

Strictly follow the Claude Code agent specification. If uncertain or spec may have changed, fetch current docs before writing any definition

Definitions live in agents/<category>/

Narrow, domain-specific agents (single language/framework, e.g. dotnet-cs-expert) live in agents/<category>/specialists/ instead of flat under the category. Broad, general-purpose agents stay flat

Agent definition style rules

  • spec compliance is critical: any broken frontmatter field, unknown key, or wrong format will silently prevent CC from loading or invoking the agent correctly; take extreme care
  • model is always inherit
  • description must be concise
  • minimize blank lines, trailing periods
  • tools list only what the agent actually needs
  • colors by category
  • no redundant concepts
  • no memory; agents run stateless
  • no bold formatting; backticks allowed where appropriate

What to check before editing

  • README.md has the full agent catalog with descriptions; keep it in sync after adding/removing agents

Skill definitions

Skills live in skills/<skill-name>/SKILL.md, one directory per skill, no category grouping. Follow the Claude Code skill specification; a skill's supporting files (scripts, references) are not version-tracked, only SKILL.md No color field — that belongs to the agent spec, not the skill spec; do not carry it over when basing a new skill on an agent or another skill

Versioning

On agent or skill definition update:

  • Regenerate .file-versions: { find agents -name "*.md" -type f; find skills -name "SKILL.md" -type f; } | sort | xargs sha256sum > .file-versions
  • In README.md, bump the agent's or skill's version per the extent of the change, and sync the parenthesized hash next to it with the first 8 chars of the new hash

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