Skip to main content

Agent Skillet

CI

Prepare and serve agent skills!

Agent Skillet helps teams install, version, and sync agent skills inside a repository.

Install

From PyPI (recommended)

uvx agent-skillet init
uv tool install agent-skillet

From source (development)

uv pip install -e .

If you do not already have uv, run:

zsh install.sh

Quick Start

  1. Initialize Skillet in your project

    skillet init
    

    This command prompts which agentic coding tool you're using, then

    • Seeds current project managed by skillet with every bundled skill in skills folder
    • Materializes them under .skillet/skills/<skill_name>/, and mirrors into configured agent folders
  2. Add new skills from Github (owner/repo or owner/repo/path[@ref]):

    skillet add anthropics/skills/frontend-design@main
    

Configuration

  1. To configure which agentic coding tool(s) skills show up for (Cursor, Claude Code, and other agents each have their own mirror paths):

    skillet config    # update your agentic coding tool
    skillet sync      # re-apply tracked sources and refresh native mirrors
    

skillet sync is a local reconciliation command. It reads .skillet/config/sources.json, materializes/removes skills under .skillet/skills/, and updates native mirror directories. It is not a "check for latest remote updates" command.

Browse skills

  1. Browse skills publicly with skillet find <query> or local ones with skillet search <query>.

How It Works

  • Tracks installed skill sources in .skillet/config/sources.json.
  • Materializes installed skills into .skillet/skills/<name>/SKILL.md.
  • Mirrors enabled skills into agent-native directories (for example .cursor/skills/ and .claude/skills/).
  • Supports local sources and GitHub specs (owner/repo, owner/repo/subpath, owner/repo/subpath@ref).
  • GitHub installs download GitHub’s full repository archive tarball for the resolved ref, then copy to the skill folder from your path segment
  • skillet sync reconciles to tracked source specs and local managed state; it should not be assumed to always pull newer upstream content when a remote ref moves.

Example sources.json

Each entry maps a skill name to its source. The kind field is either "local" or "github".

{
  "git-os": {
    "kind": "local",
    "source": "git-os"
  },
  "python-design-patterns": {
    "kind": "github",
    "source": "wshobson/agents/python-design-patterns@main"
  },
  "skill-creator": {
    "kind": "github",
    "source": "anthropics/skills/skill-creator"
  }
}

skillet add writes these entries for you — you rarely need to edit the file directly.

Common Commands Examples

# Add a local skill directory (must contain SKILL.md)
skillet add ./team-skills/checkout-flow

# Add all skills from a GitHub repo (owner/repo)
skillet add wshobson/agents

# Add a single skill from a GitHub repo (owner/repo/subpath)
skillet add anthropics/skills/skill-creator

# Pin to a specific branch or tag (owner/repo/subpath@ref)
skillet add wshobson/agents/python-design-patterns@main

# Reconcile tracked sources after editing .skillet/config/sources.json
skillet sync

# Find skills on skills.sh
skillet find <query>

# Search local skills
skillet search <query>

# List installed skills
skillet list

# Remove a skill
skillet remove skill-creator

Tip: skillet.lock records origins with a github: prefix (e.g. github:anthropics/skills/skill-creator). skillet add accepts both forms, so you can copy-paste a lock origin directly as a spec.

Bundled Skills

  • find-skills: Discover and install skills from the Agent Skillet ecosystem
  • git-os: Conventional commits, atomic changes, and GIT-OS workflow
  • sprint: Ticket-to-PR automation with branch and description templates
  • deploy-checklist: Pre/post deployment verification checklist

Documentation

Contributing

Contributions are welcome and encouraged.

  • Open an issue first for bug reports, feature requests, or design discussion.
  • Keep pull requests focused and small; include clear context in the description.
  • Add or update tests when behavior changes.
  • Run local checks before opening a PR:
uv sync
ruff check
pytest
  • Be respectful and collaborative in reviews so we can keep the project healthy and active!

License

MIT

Release files for agent-skillet 0.2.0

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Source distribution (sdist)

Source distribution for agent-skillet 0.2.0
File Size Uploaded
agent_skillet-0.2.0.tar.gz 41.6 kB Details

Built distribution (wheel)

Table of built distributions (wheels) for agent-skillet 0.2.0
File Interpreter ABI Platform
agent_skillet-0.2.0-py3-none-any.whl Python 3 none any Details

Total release size: 84.2 kB

Release files / agent_skillet-0.2.0.tar.gz

Download URL agent_skillet-0.2.0.tar.gz
Size 41.6 kB
Tags Source
SHA-256 checksum
How to use checksums
0a483898e5ed1e50fea468a7df5566e9e7502fcd64a2188658315ea37417a53a
BLAKE2b-256 checksum
How to use checksums
920472ce0186a8c1644a5fd92ca847c7104d06729619da56fcce2bbb3ac1dd2a
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/6.1.0 CPython/3.13.12

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on May 14, 2026.

Transparency log

Release files / agent_skillet-0.2.0-py3-none-any.whl

Download URL agent_skillet-0.2.0-py3-none-any.whl
Size 42.6 kB
Tags Python 3
SHA-256 checksum
How to use checksums
3ae57cd60155e92d3f1900de61d16937454730462a8023c69814093457cb6ac6
BLAKE2b-256 checksum
How to use checksums
2ab864cb781aaefeff503fe1d50ba7171a145685a7ded16b546dd12e5a1add0a
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/6.1.0 CPython/3.13.12

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on May 14, 2026.

Transparency log

Release history Release notifications | RSS feed

This release

0.2.0 This release

2 release files

0.1.1

2 release files

0.1.0

2 release files

Anthropic, PBC Visionary sponsor Bloomberg Visionary sponsor Hudson River Trading Visionary sponsor Meta Visionary sponsor NVIDIA Visionary sponsor Microsoft Sustainability sponsor Depot Continuous Integration AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page