agentskills

agentskills: Define Portable Skills for AI Agents

Agent Skills is an open format for giving AI agents specialized knowledge and repeatable workflows through portable skill folders. It uses progressive disclosure so agents load detailed instructions only when a task calls for them.

Stars
26.5k
Forks
2k
Stars, last 30 days
+1,130
Last commit
3 months ago
Contributors
41

Our take

Activeregular commits

Health score

57/100

How it is scored
Activity
35
Community
59
Issues
70
Pull requests
63

+1,130 stars in the last 30 days

This is a clearly scoped open format with recent development and broad contributor participation, though the repository has no release in the past year and lists no CI or tests in its project requirements data.

Good fit if

  • You need a portable way to package agent instructions, workflows, and supporting files.
  • Your agent or client supports the Agent Skills format, or you are evaluating how to add that support.
  • You want documentation under CC-BY-4.0 and repository code under Apache-2.0.

Look elsewhere if

  • Your target agent cannot load skills in this format and you do not plan to build compatibility.
  • You need a packaged runtime or hosted service, rather than a specification and documentation repository.
  • Your adoption process requires published releases or CI and tests in the repository, which are not indicated in the provided project data.
All health signals
Last commit2026-08-09 (3 months ago)
Commits, last 90 days14
Releases, last 12 months0
Contributors41 (top contributor: 63% of commits)
Issues closed, last 90 days16 (typically closed in 2 days)
Pull requests merged, last 90 days8 (typically merged in 0 days)
Project age10 months

Checked on 2026-10-24 with the GitHub API.

Project history

Daily GitHub snapshots since Oct 7, 2026. History covers stars and forks; health signals are not tracked historically.

Selected snapshot

Stars

27,407

Forks

2,040

Change since Nov 17, 2026: +43 stars·+1 forks

Overview

Agent Skills is an open format for extending AI agents with specialized knowledge and workflows. A skill is a folder centered on a SKILL.md file, with optional scripts, references, and assets that package procedural or team-specific context for reuse.

The format is intended for skills-compatible agents across products. Its progressive disclosure approach lets an agent discover skills from their names and descriptions, then load full instructions and supporting resources when relevant, rather than keeping every skill's details in context at all times.

Key Features

  • Defines a lightweight, folder-based format with SKILL.md as the required entry point.
  • Supports skill metadata with at least a name and description, alongside task instructions.
  • Allows optional scripts, reference materials, templates, and other resources to accompany instructions.
  • Uses discovery, activation, and execution stages to load information as needed.
  • Packages repeatable workflows and domain-specific context in version-controlled folders.
  • Is designed for reuse across compatible agent products; the project links to a client showcase.
  • Provides a specification, documentation, and example skills as starting points.

Use Cases

  • A team can package internal procedures as a version-controlled skill for compatible agents.
  • A specialist can capture a multi-step workflow, such as a review or analysis process, for more consistent reuse.
  • Developers building agent clients can consult the specification to implement support for the format.
  • Skill authors can bundle scripts and reference files alongside instructions when an agent needs those resources to complete a task.

What you need

Detected in the repository

  • Node.js (from package.json)

License in plain words

Apache-2.0permissive

  • Commercial use: yes
  • Modify and redistribute: yes
  • You must keep: the license, the NOTICE file and a note of your changes
  • Share your changes: no
  • Includes an explicit patent grant from the contributors:

A summary, not legal advice: the LICENSE file is what applies.

Getting Started

Read the documentation and specification, then explore the example skills. See the README for project details and contribution guidance.

Considerations

The repository is active, with regular commits in the measured period, and its contributors are not concentrated around one person. However, it has had no release in over a year, so teams should review the specification and its current compatibility with their target agents rather than relying on release cadence. Project data does not indicate CI or tests, and the listed runtime requirement is Node.js. Check the repository's license details when redistributing or adapting materials, since code and documentation use different licenses.

Found this useful?

Share it with someone who would like agentskills.