Helps you make better decisions when designing content schemas for headless CMSes like Sanity. The guidance is organized around specific modeling questions: should this be a reference or embedded? Is this content reusable enough to split out? Am I accidentally building page-shaped schemas instead of content-shaped ones? Instead of dumping everything at once, it points you to focused references on separation of concerns, the reuse spectrum, and taxonomy patterns. The framing is refreshingly pragmatic: content is data, not pages, and your model should work for channels that don't exist yet. Useful whether you're starting fresh or refactoring a mess.
npx skills add https://github.com/sanity-io/agent-toolkit --skill content-modeling-best-practices