awesome-ai-agent-tools
AGENTS.md62 starsChanged 3 months agoCC0-1.0View on GitHub
This file provides context for AI coding agents working with this repository. Awesome AI Agent Tools is an open-source library of installable AI agent components: skills, MCP servers, agent workflows, subagents, hooks, plugins, prompts, and tools curated from 100+ repositories. Exact per-category counts live in each catalog.json and in the generated readme.md. Do not hardcode counts in hand-written docs: they go stale. The generator reads the catalogs and emits the counts. Each directory contains a catalog.json file with structured metadata:
Where files are headingNew AGENTS.md files have more rules about keeping secrets out of the code, and fewer about code style.
Of the AGENTS.md files we read, the 31,774 first seen in the last six months, against the older ones: the kind of rule that gained most share, and the kind that lost most. About files like this in general, not this one, and an estimate, since a kind is a pattern in the words.
In this repository2 other files it gives its agents.
- loopsSkill
- awesome-ai-agent-toolsSkill
The file
We split this AGENTS.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
- Very specific: line 49
Add your entry to the correct catalog.json
- Very specific: line 50
Ensure JSON is valid: cat skills/catalog.json | python3 -m json.tool
- Distinctive: line 59
All entries are validated by GitHub Actions on PR
≈ “Common” is an estimate: at least 1 in 5 of the 40,246 AGENTS.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
AGENTS.md
This file provides context for AI coding agents working with this repository.
What Is This Repo?
Awesome AI Agent Tools is an open-source library of installable AI agent components: skills, MCP servers, agent workflows, subagents, hooks, plugins, prompts, and tools curated from 100+ repositories.
Repository Structure
| Directory | What It Contains |
| skills/ | SKILL.md files: reusable instruction sets that teach AI agents new capabilities |
| mcps/ | Model Context Protocol server configs with install commands |
| loops/ | Agent workflow patterns with verification criteria |
| subagents/ | Specialized agent definitions with model routing |
| hooks/ | Claude Code hooks for security, automation, and quality |
| plugins/ | Extensions for Claude Code, OpenCode, Cursor, and 6 more platforms |
| prompts/ | Curated prompt collections and marketplace links |
| tools/ | CLI utilities that enhance agent capabilities |
| mods/ | Claude Code mods: plugins that customize the Claude Code experience |
Exact per-category counts live in each catalog.json and in the generated readme.md. Do not hardcode counts in hand-written docs: they go stale. The generator reads the catalogs and emits the counts.
Catalog Format
Each directory contains a catalog.json file with structured metadata:
{
"skills": [
{
"id": "skill-name",
"name": "Display Name",
"category": "Development",
"description": "What this skill does",
"githubUrl": "https://github.com/owner/repo",
"installCommand": "npx skills add owner/repo --skill skill-name"
}
]
}Claude Code mod entries also require source (owner/repo), installCommand, compatibility, and license. Mods execute with Claude Code's available access; listing one does not mean it has passed a security audit.
How to Contribute
- Fork the repo
- Add your entry to the correct catalog.json
- Ensure JSON is valid:
cat skills/catalog.json | python3 -m json.tool - Submit a PR
See contributing.md for full details. It is also the agent contribution skill: give it to your AI agent and it handles everything.
Standards
- Skills follow the SKILL.md open standard
- MCP servers follow the Model Context Protocol
- All entries are validated by GitHub Actions on PR
Key Files
readme.md- Main overview and catalogSKILL.md- Agent skill for installing from this collectioncontributing.md- Human and agent contribution guideskills/catalog.json- Machine-readable skills catalogmcps/catalog.json- Machine-readable MCP catalogloops/catalog.json- Machine-readable loops catalogmods/catalog.json- Machine-readable Claude Code mods catalog
Review policy: one human approval from someone other than the PR author, plus a passing catalog validation check. An approval should say “Go” and give one brief reason. The repository owner must configure this approval count in GitHub branch protection; documentation alone does not enforce it.
Claude Code mods are executable plugins and run with the user's Claude Code access. Require clear source attribution and available install, compatibility, and license details; do not describe listing as a security audit.
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.