Hooks in Claude Code

What are hooks?

Hooks are scripts that run automatically when Claude Code performs specific actions. They can validate, log, or transform tool calls.

Hook types

Hook	Runs when
`PreToolUse`	Before Claude calls a tool (can block it)
`PostToolUse`	After a tool call completes
`Notification`	When Claude sends a notification
`Stop`	When Claude finishes a turn

Configuration

Add hooks to .claude/settings.json:

{
  "hooks": {
    "PreToolUse": [
      {
        "matcher": "Write",
        "hooks": [{
          "type": "command",
          "command": "python3 .claude/hooks/validate.py"
        }]
      }
    ]
  }
}

Example: block writes to protected files

import sys, json

input_data = json.load(sys.stdin)
file_path = input_data.get("tool_input", {}).get("file_path", "")

protected = [".env", "secrets.json", "credentials.yml"]
if any(file_path.endswith(p) for p in protected):
    print(json.dumps({"decision": "block", "reason": "Protected file"}))
else:
    print(json.dumps({"decision": "approve"}))

Use cases

Block writes to sensitive files
Auto-format code after edits
Log all tool calls for audit
Run linters after file changes

Source: docs.anthropic.com/en/docs/claude-code/hooks