
If you're building skills for Claude, this is your implementation guide. It walks through the three-level loading system (metadata, SKILL.md body, bundled resources), explains when to use scripts versus reference docs versus assets, and hammers home the core principle: be concise because context window is shared space. The progressive disclosure patterns are especially useful, showing you exactly how to keep SKILL.md under 500 lines by splitting variants and detailed specs into separate files. Think of it as the spec for the spec. Most valuable when you're past your first skill and wondering why Claude isn't picking up your carefully written instructions.
$ npx -y skills add composiohq/skills --skill skill-creator --agent claude-codeInstalls into .claude/skills of the current project.
This skill provides guidance for creating effective skills.
Skills are modular, self-contained packages that extend Claude's capabilities by providing specialized knowledge, workflows, and tools. Think of them as "onboarding guides" for specific domains or tasks—they transform Claude from a general-purpose agent into a specialized agent equipped with procedural knowledge that no model can fully possess.
The context window is a public good. Skills share the context window with everything else Claude needs: system prompt, conversation history, other Skills' metadata, and the actual user request.
Sponsors