# agentskills/agentskills

Specification and documentation for Agent Skills

Repository: https://github.com/agentskills/agentskills
Canonical: https://ross.abutalabs.com/products/agentskills
Homepage: https://agentskills.io
Language: Python
License: Apache-2.0
License Family: permissive
Topics: agent-skills
Last push: 2026-08-09T20:36:04+00:00

## Health v2 (maintenance only)
Score: 59/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 96, release rhythm 35, longevity 18
- inputs: {"age_days": 260, "days_push": 24, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 24748, forks 1835 (observed 2026-08-28T04:11:37.635576+00:00)

## What it is
Agent Skills is an open specification and documentation for a lightweight, portable format that extends AI agent capabilities using folders containing a SKILL.md file with metadata and instructions. Skills bundle procedural knowledge, scripts, and resources that agents load on demand via progressive disclosure across any compatible agent client.

## Use cases
- create reusable skills for AI coding agents
- package team workflows so agents follow consistent procedures
- give an AI agent domain expertise like legal review or data analysis
- share skills across Claude Code, VS Code Copilot, and OpenAI Codex
- reduce agent context usage with progressive skill loading
- version-control agent instructions and templates

## When to choose
- you want a vendor-neutral, open standard for extending agent capabilities
- you need portable, version-controlled procedural knowledge usable across multiple agent products
- you want skills to load on demand with minimal context footprint

## When to avoid
- you need a runtime or execution engine rather than a format specification
- your agent platform does not support the Agent Skills format
- you want fully autonomous agent orchestration rather than on-demand instruction loading

## Facets
- artifact type: learning-resource
- maturity: active
- function: agent-framework, prompt-engineering, documentation
- domain: large-language-models, developer-tools
- platform: cross-platform, python
- tags: agent-skills, open-standard, specification, skill-format, progressive-disclosure, skyll-md, anthropic, ai-agents, automation

## Member repositories
- agentskills/agentskills (main) score 59

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:11:37.635576+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-29T16:56:09.741045+00:00, confidence not recorded.
  - readme: https://github.com/agentskills/agentskills (fetched 2026-08-28T04:11:37.635576+00:00, sha b25541108198)
  - homepage: https://agentskills.io (fetched 2026-08-29T07:53:22.901359+00:00, sha 7aea606585f9)
  - site_page: https://agentskills.io/skill-creation/quickstart (fetched 2026-08-29T07:53:22.911417+00:00, sha 2f3334527656)
- Data as of 2026-08-30T08:39:29.467469+00:00.