This is meta tooling for building and refining skills themselves. You walk through capturing intent, drafting the skill markdown, writing test cases, running evals, and iterating based on results. It includes a separate workflow for optimizing skill descriptions to improve triggering accuracy, which matters since Claude tends to undertrigger skills. The progressive disclosure guidance is solid: keep metadata tight, SKILL.md under 500 lines, and push detailed references into separate files. Most useful when you're systematically developing a skill rather than just hacking something together, though it adapts if you want to skip the eval rigor and just iterate by feel.
npx -y skills add ar9av/obsidian-wiki --skill skill-creator --agent claude-codeInstalls into .claude/skills of the current project.
Select a file.
supercent-io/skills-template
supercent-io/skills-template
huangjia2019/claude-code-engineering
reactjs/react.dev
reactjs/react.dev