If you're building skills for Claude Code or other AI coding assistants, this walks you through the entire Agent Skills spec with practical structure advice. It covers frontmatter requirements, progressive disclosure patterns, and when to split content into reference files. The security section is especially thorough, spelling out when and how to surface risks so the LLM actually catches them. Useful if you're maintaining a skill library and want consistency, or if you're writing your first SKILL.md and don't want to guess at field names and file organization. Written by the Apollo team, so the examples lean GraphQL but the structure guidance applies broadly.
npx skills add https://github.com/apollographql/skills --skill skill-creator