agent-skills
CLAUDE.md2 starsChanged 3 days agoMITView on GitHub
This repository contains AI agent skills in the skills.sh format, installed via npx skills add. Claude Code exposes every skill as a slash command, so anything meant to run from /name is a hand-invoked skill (step 2 below), not a .claude/commands/ file. Skills follow the skills.sh format and come in two shapes:
In this repository11 other files it gives its agents.
- skills/feature-workflow/AGENTS.mdAGENTS.md
- skills/react-query-best-practices/AGENTS.mdAGENTS.md
- smc-crt-analysisAGENTS.md
- skills/validate-pr/AGENTS.mdAGENTS.md
- babysit-prSkill
- feature-workflowSkill
- qa-verifySkill
- react-query-best-practicesSkill
- react-query-migrate-v5Skill
- smc-crt-analysisSkill
- validate-prSkill
The file
We split this CLAUDE.md into the rules it gives an agent
3 rules. About 0 are common, 2 are distinctive, and 1 is 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.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 33
Create skills/<skill-name>/ with a SKILL.md whose YAML frontmatter has name, description, license: MIT, and metadata (author, version).
- Distinctive: line 35
Keep skills project-agnostic: this repo is public.
- Distinctive: line 37
Add it to README.md: a section and an install line.
≈ “Common” is an estimate: at least 1 in 5 of the 27,641 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.
Files
CLAUDE.md
This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository.
Repository Purpose
This repository contains AI agent skills in the skills.sh format, installed via npx skills add. Claude Code exposes every skill as a slash command, so anything meant to run from /name is a hand-invoked skill (step 2 below), not a .claude/commands/ file.
Repository Structure
skills/ # skills.sh format skills (source of truth)
<skill-name>/
SKILL.md # Instructions for the agent (frontmatter + content)
AGENTS.md # Rule-based skills: compiled document, all rules expanded
rules/ # Rule-based skills: one file per rule
<topic>.md # Workflow skills: reference disclosed from SKILL.md
agents/openai.yaml # Optional: Codex display name and invocation policy
.agents/skills/<skill-name> # Symlink -> ../../skills/<skill-name>
.claude/skills/<skill-name> # Symlink -> ../../skills/<skill-name>Creating New Skills (skills.sh format)
Skills follow the skills.sh format and come in two shapes:
- Rule-based (
react-query-best-practices,validate-pr): a body of guidance the agent consults.SKILL.mdindexes the rules,rules/holds one file per rule (brief explanation + incorrect/correct examples), andAGENTS.mdcompiles them all for distribution. Any change to a rule file updates theSKILL.mdindex andAGENTS.mdin the same commit. Forreact-query-best-practices, rebuild withscripts/build-agents-md.py skills/react-query-best-practices(--checkverifies index, rules andAGENTS.mdagree). - Workflow (
qa-verify,babysit-pr,react-query-migrate-v5): ordered steps the agent executes, each ending on a checkable completion criterion.SKILL.mdholds the steps; reference that only some branches need moves to a sibling<topic>.mdlinked from the step that uses it. Norules/orAGENTS.md.
To add one:
- Create
skills/<skill-name>/with aSKILL.mdwhose YAML frontmatter hasname,description,license: MIT, andmetadata(author,version). - Pick the invocation: keep a trigger-rich
descriptionwhen the agent should fire the skill on its own; setdisable-model-invocation: true(and a one-line, human-facingdescription) when it only runs by hand. Mirror a hand-run skill inagents/openai.yamlwithallow_implicit_invocation: false. - Keep skills project-agnostic: this repo is public. Read project specifics (commands, logins, channels) from the target repo's
AGENTS.md/CLAUDE.mdinstead of hard-coding them. - Symlink it:
ln -s ../../skills/<skill-name> .agents/skills/<skill-name>and the same under.claude/skills/. - Add it to
README.md: a section and an install line.
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.

