/hook-creator

The /hook-creator skill teaches you how to create and configure Claude Code hooks for event-driven automation.

What are Hooks?

Hooks are shell commands or LLM-evaluated prompts that execute in response to Claude Code events. They provide programmatic control over Claude's behavior without modifying core code.

Quick Start Workflow

1. Create hooks config file:
   • Project: .claude/hooks.json
   • User: ~/.claude/hooks.json
2. Choose hook event (when it fires)
3. Choose hook type (command or prompt)
4. Configure matcher (which tools trigger it)
5. Test with claude --debug

Example: Log All Bash Commands

.claude/hooks.json:

{
  "hooks": {
    "PreToolUse": [
      {
        "matcher": "Bash",
        "hooks": [
          {
            "type": "command",
            "command": "jq -r '.tool_input.command' >> ~/.claude/bash-log.txt"
          }
        ]
      }
    ]
  }
}

This hook:

• Fires before (PreToolUse) every Bash tool use
• Executes a command (not an LLM prompt)
• Logs the command to a file

Hook Types

Hook Types

Command Hooks

Execute shell commands:

{
  "type": "command",
  "command": "echo 'Tool used' >> log.txt"
}

Prompt Hooks

Use LLM to evaluate:

{
  "type": "prompt",
  "prompt": "Is this command safe to run?"
}

Blocking Hooks

Blocking hooks can return "decision": "block" to prevent actions:

{
  "type": "command",
  "command": "if [[ \"$COMMAND\" == *\"rm -rf\"* ]]; then echo '{\"decision\": \"block\", \"reason\": \"Dangerous command\"}'; fi"
}

Use Cases

1. Logging - Track all tool usage
2. Validation - Block dangerous commands
3. Notifications - Alert on specific events
4. Context injection - Add information before tools run
5. Workflow automation - Trigger actions after completion