📄 中文摘要
Claude Code 钩子允许在代理的生命周期关键点运行脚本，包括工具使用前、使用后、停止时和出错时。钩子是 Claude Code 在生命周期事件中调用的可执行脚本，具有多种功能，如阻止某个操作、记录事件、执行副作用（如格式化代码、发送通知）以及将上下文注入代理。钩子配置存储在 .claude/settings.json 文件的 hooks 键下。文章中列出了五个常用的钩子，帮助开发者在项目中有效管理和监控工具的使用。
我在每个项目中使用的 5 个 Claude Code 钩子
出处: 5 Claude Code hooks I use on every project
发布: 2026年3月17日
📄 English Summary
5 Claude Code hooks I use on every project
Claude Code hooks allow scripts to run at key points in the agent's lifecycle, including before and after tool use, on stop, and on error. Hooks are executable scripts called by Claude Code during lifecycle events, serving various purposes such as blocking actions, logging events, performing side effects (like formatting code or sending notifications), and injecting context back into the agent. The configuration for hooks is stored under the hooks key in the .claude/settings.json file. Five commonly used hooks are discussed, aiding developers in effectively managing and monitoring tool usage in their projects.