This workflow turns websites or topics into organized markdown that LLMs can actually work with. Point it at documentation sites, API references, or even just a topic, and it'll use Firecrawl's map and scrape tools to pull everything down into a structured knowledge base. The onboarding is refreshingly sparse,it infers what you want and only asks clarifying questions if truly blocked. You get different output modes depending on whether you're building RAG chunks, fine-tuning datasets, or just mirroring docs locally. The parallel execution for large sites and the opinionated file structure make it feel production-ready rather than proof-of-concept. Good fit if you're tired of manually organizing web content for AI context.
npx -y skills add firecrawl/firecrawl-workflows --skill firecrawl-knowledge-base --agent claude-codeInstalls into .claude/skills of the current project.
Select a file.
juliusbrussee/caveman
mattpocock/skills
shadcn/improve
obra/superpowers
forrestchang/andrej-karpathy-skills
vercel-labs/skills