复制下面这句话,粘贴给 Claude Code、Codex、Cursor 等 AI 编程工具,它会读取安装说明并在你确认后完成安装。
请阅读 https://ai.atlankj.com/install/asset/gh-claude-skills-5c86291fcd9d ,按照其中的说明把「agent-launcher-skills Hooks · hooks」安装到你(当前 AI 工具)中。执行前先告诉我将运行的命令和写入的位置,等我确认。
查看 AI 将读取的安装说明正在读取 GitHub 原文…
内容来自 GitHub 原始文件,由原作者维护。在 GitHub 查看
{
"description": "Opt-in SessionStart/SessionEnd hooks for agent-launcher. BOTH are disabled unless AGENT_LAUNCHER_SESSION=1 is set, so they never fire in unrelated repos. SessionStart surfaces the current CMA launch goal (./my-agent/goal.json) as <agent_launcher_goal> data so a multi-session launch resumes where it stopped. SessionEnd reminds you to checkpoint/advance the goal. Any error exits 0 — a hook can never break a session. Additionally disable SessionEnd with AGENT_LAUNCHER_SESSIONEND=0.",
"hooks": {
"SessionStart": [
{
"hooks": [
{
"type": "command",
"command": "python3 ${CLAUDE_PLUGIN_ROOT}/hooks/session_start.py"
}
]
}
],
"SessionEnd": [
{
"hooks": [
{
"type": "command",
"command": "python3 ${CLAUDE_PLUGIN_ROOT}/hooks/session_end.py"
}
]
}
]
}
}