agentskills: Standardize Skills for AI Agents
Summary
Agent Skills defines an open format for packaging instructions, resources, and optional scripts that give AI agents specialized capabilities. Use it to create portable, reusable workflows for compatible agent tools.
At a glance
- Language
- Python
- License
- Apache-2.0
- Stars
- 25.9k
- Forks
- 1.9k
- Added to OSRepos
- August 18, 2026
- Last analyzed
- October 3, 2026
Topics
Click on any tag to explore related repositories
Use at your own risk
OSRepos shares public repositories for knowledge and discovery only. Any installation, execution, configuration, or use of code from these repositories is the user's own responsibility. Always review the repository, source code, dependencies, licenses, and security implications before running or installing anything. OSRepos is not responsible for issues, damages, or losses resulting from third-party repositories.
Overview
Agent Skills is an open specification and documentation project for adding task-specific knowledge and workflows to AI agents. It addresses the challenge of giving agents reliable, reusable context without loading every instruction into their working context at once.
The format is intended for skill authors and developers of compatible agent products. It packages a skill as a folder with a SKILL.md file and optional supporting resources, using progressive disclosure so an agent can discover a skill before loading its full instructions. See the documentation and specification for the canonical format details.
Key Features
- Defines a lightweight, open format for describing agent skills.
- Requires a
SKILL.mdfile with a skill name and description, at minimum. - Supports optional scripts, reference materials, templates, and other resources alongside instructions.
- Uses progressive disclosure: agents discover skills by metadata, then load full instructions when relevant.
- Supports reusable, version-controlled procedures and domain-specific context.
- Is designed for use across compatible agent products, rather than a single client.
Use Cases
- An agent developer can add a consistent skill-loading format to a product that needs to expose task-specific procedures.
- A team can package internal workflows, such as review steps or data-analysis procedures, for compatible agents to use repeatedly.
- A skill author can share domain knowledge and supporting files as a portable folder instead of embedding all instructions in an application.
- A user of multiple compatible agent tools can maintain skills in a format intended to work across products.
Project Facts
- Language: Python
- License: Apache-2.0
- Stars: 25.9k
- Forks: 1.9k
- Topics: agent-skills
- Archived: No
Getting Started
Start with the documentation and specification to create or integrate skills. This repository provides the format and its documentation, not a setup command; see the README for project details.
Considerations
- This repository specifies a format; using skills requires an agent product that supports the format.
- The README does not provide an installation command or a bundled agent runtime.
- The repository lists Apache 2.0 for code and CC-BY-4.0 for documentation, with directory-specific details noted in the repository.
Source repository
Open the original repository on GitHub.
18 counted GitHub visits
Related repositories
Similar repositories that may be relevant next.
skills: A Collection of Agent Skills for Building Great Interfaces
September 22, 2026
The `skills` repository by Jakub Krehel offers a comprehensive collection of agent skills designed to enhance interface development. It provides tools and guidelines covering UI, typography, colors, accessibility, layout, and product writing. This resource is invaluable for developers and designers aiming to create high-quality, user-friendly interfaces.
karpathy-llm-wiki: Build a Karpathy-Style LLM Knowledge Base
September 9, 2026
karpathy-llm-wiki is an Agent Skills-compatible tool that implements Karpathy's LLM Wiki idea, allowing users to build a durable knowledge base. It enables LLMs to maintain structured wiki pages by ingesting sources, compiling knowledge, and answering questions with citations. This project offers a robust alternative to traditional RAG for compounding knowledge over time.
best-skills: Compare Agent Skill Rankings Across Platforms
September 8, 2026
Best Skills publishes daily rankings of agent skills using install counts, growth, maintenance, publisher verification, and social activity across several ecosystems. Its CSV data helps developers compare skills without treating platform-specific counts as directly equivalent.
context-engineering-kit: Improve AI Coding Agent Workflows
September 6, 2026
A collection of installable skills, commands, and agents for improving coding-agent reliability and code quality. Use it to add reflection, specification-driven development, multi-agent workflows, and code review to supported AI coding tools.