cc-hook-test

Test runner for Claude Code hooks. Auto-detects hook type and generates test cases.

npx cc-hook-test ~/.claude/hooks/destructive-guard.sh

  cc-hook-test — destructive-guard.sh
  Detected type: PreToolUse

  PASS empty input passes
  PASS safe command passes
  PASS blocks rm -rf /
  PASS blocks rm -rf ~
  PASS allows rm -rf node_modules
  PASS blocks git reset --hard

  6/6 tests passed

Usage

# Auto-test a hook (generates test cases based on script content)
npx cc-hook-test <hook-script>

# Test with custom test cases
npx cc-hook-test <hook-script> tests.json

# Generate a test template
npx cc-hook-test --generate destructive
npx cc-hook-test --generate branch
npx cc-hook-test --generate secret
npx cc-hook-test --generate custom

Test File Format

[
  {
    "name": "blocks rm -rf /",
    "input": { "tool_input": { "command": "rm -rf /" } },
    "expect_exit": 2,
    "expect_stderr": "BLOCKED"
  },
  {
    "name": "allows ls",
    "input": { "tool_input": { "command": "ls -la" } },
    "expect_exit": 0
  }
]

What It Auto-Detects

| Hook Content | Detected Type | Generated Tests | |---|---|---| | rm -rf, git reset --hard | PreToolUse (destructive) | rm variants, git reset, git clean | | git push, main, master | PreToolUse (branch) | push to main, force push, feature push | | .env, secret, credential | PreToolUse (secret) | git add .env, git add src/ | | py_compile, file_path | PostToolUse (syntax) | non-existent file | | stop_reason | Stop | normal/user stop |

Supports

• Bash hooks (.sh)
• Node.js hooks (.js)
• Python hooks (.py)

Related

• cc-safe-setup — install safety hooks with one command
• COOKBOOK.md — 19 hook recipes

License

MIT