What problem does it solve?
This Skill helps you create and implement Claude Code plugin hooks that safely validate and control tool usage across important workflow events, reducing errors and preventing unsafe operations.
Core Features & Use Cases
- Prompt-based hook design: Create context-aware validations that can approve, deny, ask, or block tool actions based on event inputs.
- Command hook guidance: Implement deterministic checks (e.g., schema validation, filesystem checks) using portable paths via ${CLAUDE_PLUGIN_ROOT}.
- Correct configuration & output contracts: Use the right hooks.json vs settings.json formats, and produce the expected structured JSON responses and exit-code behavior.
- Lifecycle coverage: Apply hooks to events like PreToolUse, PostToolUse, Stop, SessionStart, PreCompact, Notification, and more.
- Testing and hardening workflow: Use the provided utility scripts (schema validation, hook testing, and linting) to reduce deployment risk.
Core Features & Use Cases
Use case example: You want to prevent dangerous file writes and require completion checks before stopping—use PreToolUse to validate Write/Edit operations, and use Stop to verify tests/build steps from the transcript.
Quick Start
Ask your AI assistant to “create a PreToolUse prompt-based hook that validates Write|Edit tool calls for path traversal and secret-file writes, then provide the exact hooks.json configuration format and the required hook output JSON contract.”
