Can hooks prevent a tool from running?

Yes, hooks triggered by PreToolUse can be used for validation and blocking, allowing you to intercept and manage tool execution based on custom logic.

Can I use regex for tool matchers?

Yes, matchers support exact string matches (e.g., 'Write') and regex patterns (e.g., 'Edit|Write' or 'Notebook.*') to trigger hooks for specific tools.

What information do hooks receive when they run?

Hooks receive a JSON object via stdin containing session metadata, the current working directory, permission modes, and event-specific data like tool inputs or responses.

What are Claude Code hooks?

Hooks are custom commands or LLM prompts that execute automatically in response to specific Claude Code events, such as tool usage, session starts, or user prompt submissions.

Where are hook configurations stored?

Hooks can be stored in multiple locations including managed-settings.json, .claude/settings.json (local or project), plugin folders, or directly in skill frontmatter.

Claude Hooks Reference 2026

Name: Claude Hooks Reference 2026
Author: Jamie-BitFlight

byJamie-BitFlight

•

Developer Tools

Provides a comprehensive technical reference for the Claude Code hooks system to enable advanced workflow automation and tool validation.

This skill acts as a definitive documentation resource for the Claude Code hooks architecture as of January 2026. It guides developers through the entire lifecycle of hook events—from tool execution triggers like PreToolUse and PostToolUse to session-level events like SessionStart and SubagentStop. By providing detailed schemas for JSON inputs, configuration precedence rules, and regex-based matcher syntax, this reference empowers users to build sophisticated automation, security policies, and custom formatting scripts that integrate seamlessly with Claude's internal operations.

Key Features

01Detailed configuration hierarchy for managed, project, user, and plugin settings.

02Complete documentation of all hook events including tool, session, and subagent triggers.

03Regex and tool-specific matcher patterns for granular execution control.

0418 GitHub stars

05Guidance on implementing scoped hooks within Skills, Agents, and Slash commands.

06Full JSON input schemas for stdin communication between Claude and hook scripts.

Use Cases

01Enforcing code style and linting automatically after Claude performs file edits.

02Initializing complex repository environments and environment variables during session startup.

03Setting up automated security audits and permission blocking before sensitive bash commands execute.

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add jamie-bitflight/claude_skills claude-hooks-reference-2026

For use in Claude.ai and ChatGPT

Download Skill