Agentic Awesome Skills (AAS Core)
A 2,100+ skill catalog plus AAS Core, a local CLI/MCP control plane that lets Codex or Claude search, compose and validate an exact, reviewable skill stack before anything is installed.
Overview
Agentic Awesome Skills is one of the largest open community catalogs of agent
skills (2,100+ entries across development, testing, security, infrastructure,
product and marketing). Its real product, AAS Core, is a local, read-only
MCP server plus an aas CLI: instead of dumping the whole catalog into your
project, it lets your coding agent search the full catalog itself, choose exact
skill IDs, and validate that selection into a reviewable, versioned
aas-stack.json before any file is written.
How it works
- Search, don't install everything. The MCP exposes
search_skills,get_skill,compose_stackand related tools so the agent inspects the catalog without you cloning thousands of files. - Agent-owned selection. The agent (not AAS) picks the exact skill IDs relevant to your project's surface (architecture, security, testing, etc.).
- Validate, then plan.
aas stack validateandaas stack plancheck the proposal and produce an immutable install plan, without applying any changes. - Install narrowly. Run the direct installer with
--skills id1,id2(or a specialized plugin/bundle) for a scoped, reviewable install; avoid--all.
Licensing note
The AAS Core tooling itself is MIT. The individual skills inside skills/
are contributed from hundreds of separate upstream repos, each under its own
license (MIT, Apache-2.0, AGPL-3.0, CC-BY-4.0 and others). Check the source
repo's own attribution ledger before reusing a specific skill's content, rather
than assuming the root MIT license applies to everything.
Examples
"Search the AAS catalog for a code-review skill for a Next.js + Postgres repo,
and propose an exact stack for me to review before installing anything."
Best used for targeted discovery (finding one good skill for a specific gap in your project) rather than as a one-shot full-catalog install.
Installation
You need Node.js and a coding agent that supports MCP (Claude Code, Cursor, Codex CLI, Gemini CLI).
- Configure the AAS MCP server in your agent so it can search the catalog (see the project's README for the exact MCP config block).
- Ask your agent to search the catalog for what you need. It returns specific skill IDs instead of the whole 2,100+ entry list.
- Preview a dry run of just those IDs:
npx agentic-awesome-skills --skills <skill-id-1>,<skill-id-2> --dry-run - Once you're happy with the plan, run the same command without
--dry-runto install only those skills. Avoid--all.
Related assets
Frontend Design
Generates distinctive, production-grade UI that avoids generic "AI slop," with bold typography, cohesive color, and precise motion, implemented as real code.
Langflow
Visual Python platform for building and deploying AI agents and workflows, with full component source access, multi-agent orchestration and one-click export as an API or MCP server.
Graphify
Turns a codebase, plus its docs, PDFs, images, and video, into a queryable knowledge graph that a coding agent can traverse instead of grepping raw files, and never guesses silently: every edge is tagged EXTRACTED, INFERRED, or AMBIGUOUS so you know what was found versus inferred.