This is a semantic search wrapper around Exa's API that gives you embeddings-based search, similar page discovery, and structured research queries. You get five endpoints: search (with neural, fast, auto, or deep modes), contents extraction, find similar by URL, direct answer, and research with custom output schemas. It's built with a two-phase architecture where the main skill handles query construction and a sub-skill executes the HTTP calls to avoid token bloat. Nice touches include category filters (research papers, GitHub repos, tweets), date ranges, domain filtering, and cost tracking per request. If you need better-than-Google semantic search or want to find content similar to a given URL, this does the job without much ceremony.
npx -y skills add benedictking/exa-search --skill exa-search --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