This is the complete technical reference for Claude Code's hook system. If you're writing hooks, you need this. It documents all 11 hook types with exact input/output JSON schemas, explains which hooks can block execution versus just inject context, and covers registration patterns in settings.json. The critical details are here: PostToolUse uses tool_response not tool_result, Stop hooks require checking stop_hook_active to avoid infinite loops, and SessionStart can persist environment variables through CLAUDE_ENV_FILE. It's organized as a quick reference table plus detailed schemas for each hook type. Bookmark this if you're doing any serious hook development, it'll save you from guessing at field names.
npx -y skills add parcadei/continuous-claude-v3 --skill hook-developer --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