anthropics/skills
Table of content
Anthropic’s public skills repository. the official catalog of agent skills for Claude Code and compatible agents.
why it matters
when Anthropic open-sources their skills catalog, it’s a signal: skills are infrastructure now.
this isn’t a side project or community experiment. this is Anthropic saying “here’s how we think agents should work” and handing you the blueprints. skills are shareable, versioned, community-maintained primitives. your agent doesn’t just run tools — it inherits an entire ecosystem of verified behaviors.
AGENTS.md is eating the world, one skill at a time.
what’s inside
verified, production-ready skills for:
- coding workflows
- context engineering
- tool integration
- multi-step automation
- agent collaboration
these aren’t toy examples. these are the skills Anthropic uses internally, now public and forkable.
the ecosystem play
Anthropic and Hugging Face both dropped public skills repos on the same day (February 26, 2026). not a coincidence.
when the two leading players in AI agents open-source their catalogs simultaneously, it’s ecosystem consolidation. skills are becoming the new apps. and like apps, they need registries, verification, versioning.
what this enables
- shareable agent behaviors across teams
- community-maintained skill libraries
- verified, audited workflows
- skills as composable primitives
- agent capabilities you can fork and customize
if your personal AI OS is a repo, skills are the packages you import.