Create event-driven Claude Code hooks for tool-use monitoring and validation.
npx skills add https://github.com/kvnloo/evolve --skill create-hooks-kvnloo
Please help me install this Agent Skill. Skill: create-hooks Source: https://github.com/kvnloo/evolve/tree/main/.claude/skills/create-hooks Command: npx skills add https://github.com/kvnloo/evolve --skill create-hooks-kvnloo
This Skill provides expert guidance to create, configure, and debug Claude Code hooks, enabling programmatic control over Claude's behavior without modifying core code. It eliminates manual intervention for repetitive tasks, validates commands, and injects context, saving you time and reducing errors in your development workflows.
prettier after every Write or Edit tool use, ensuring consistent code style across your project without any manual effort.Create a .claude/hooks.json file in your project.
Add a PreToolUse hook with a Bash matcher.
Configure it to log every bash command to ~/.claude/bash-log.txt.
Test your new hook with claude --debug.
High-intent search queries and answers about installing and using this skill.
Event-driven hooks monitor Claude Code tool usage and trigger automated actions on events like PreToolUse, PostToolUse, or SessionStart. Create a `.claude/hooks.json` file with hook configurations that execute bash commands or LLM prompts in response to these events, enabling automation without modifying core code.
Yes. Use PreToolUse hooks with Bash matchers to validate commands before execution. Hooks support LLM reasoning to inspect tool inputs and prevent destructive actions, or return stop signals to block execution entirely.
Claude Code supports PreToolUse, PostToolUse, Stop, SessionStart, and UserPromptSubmit events across Bash, Write/Edit, and MCP tools. Each event fires at specific workflow moments, letting you inject context, log activity, validate inputs, or trigger notifications.
Configure a SessionStart hook to execute LLM prompts that load project-specific context into Claude's reasoning. This prepares Claude with environment details, coding standards, or domain knowledge without manual input on each session.
Yes. Use a PostToolUse hook with a Write/Edit matcher to trigger code formatting tools like Prettier after every file modification. This enforces consistent code style across your project without manual intervention.
Command hooks execute shell scripts synchronously and return output; prompt hooks submit LLM reasoning tasks that inspect tool inputs or context and return structured decisions like validation results or stop signals for advanced control.
89jobrienConfigure Claude Code hooks for workflow automation and audit logging.
tldrawAuthor Playwright-based end-to-end tests for tldraw UI interactions.
yu-iskwImplement and validate Claude Code hook configurations and executable scripts.
xiaoyuge886Create event-driven hooks for Claude Code plugin validation and security policies.