Documentation

Everything you need to understand agent skills, catalogs, validation, and how to build reliable AI agent capabilities.

About Skills

Understand the fundamentals of agent skills and how they work

Learn about SKILL.md files, how they extend agent capabilities, and how skills are loaded and invoked.

Understand the role of agent harnesses—long-running processes that execute skills across different AI platforms.

Safety concerns, harness compatibility, token consumption, and how progressive disclosure helps.

Compare skills to MCP servers and sub-agents. Understand when to use each approach.

Discover, publish, and manage skills across catalogs

Explore public, organization, and personal catalogs. Learn about publishing flows and namespacing.

Curated skill bundles that work together

Learn about base packs, editions, shared context, and how packs accelerate your workflow.

Clone pack skills to your personal catalog with shared context merged automatically.

Build your own packs or fork existing ones. Add domain context and share with the community.

Quality assurance and certification for skills

Understand the validation pipeline: safety scans, quality grading, and certification process.

Create, edit, and test your own skills

Create skills with our editor. Generate with AI, fork from catalog, or write from scratch.

Official toolkits and utilities for skill development

Our official skill pack for building, validating, testing, and publishing skills.

Authenticate skills that connect to the catalog API. One key, saved once, works everywhere.

The open standard for portable agent skills