القائمة
创建和管理 Agent Teams 并行任务团队。当用户说"建个团队"、"拉个队伍"、"并行跑"、"create team"、"spawn agents"、"组队"时触发。
التثبيت باستخدام Codex أو Claude انسخ هذا Prompt والصقه في Codex أو Claude أو مساعد آخر ليراجع صفحة Skill ويثبّتها لك.
استنادا إلى تصنيف SOC المهني
This skill should be used when users need to install, configure, or troubleshoot DeepEP (DeepSeek Expert Parallelism) on NVIDIA GPU systems. It covers the complete installation workflow including CUDA, DOCA-OFED, NVSHMEM with IBGDA support, and DeepEP itself. The skill is particularly useful for B200/H100/A100 GPUs with RoCE/InfiniBand networking, and includes comprehensive debugging capabilities for common installation failures.
查看 Agent Teams 状态。当用户说"teams"、"团队状态"、"看看团队"、"team status"时触发。
Generate videos using Google Veo 3.1 on Vertex AI. Use when the user says "生成视频", "做个视频", "generate video", "帮我生成一段视频", "create a video", "视频生成", "text to video", "文生视频", or when you need to create video content.
Generate images using Google Imagen 4 on Vertex AI. Use when the user says "生成图片", "画一张", "generate image", "帮我画", "生成一张图", "create an image", "图片生成", or when you need to create visual content to explain concepts.
Deploy a Discord Bot that connects to Claude Code via persistent processes using Unix socketpair + stream-json (same mechanism as VSCode extension). Each user gets their own long-running Claude process with full interactive mode support (auto memory, CLAUDE.md, skills). Includes Whisper voice transcription, user whitelist, session history with dropdown switcher, and slash commands. Use when the user says "帮我建一个 Discord Bot", "setup discord bot", "搭建 Discord 机器人", "discord bot 设置", or "部署 discord bot".
安装配置 nginx 静态文件服务器，用于展示 Claude Code 生成的富内容（表格、图表、报告等）。当用户说"装nginx"、"配web服务"、"setup web server"时触发。
| name | agent-teams |
| description | 创建和管理 Agent Teams 并行任务团队。当用户说"建个团队"、"拉个队伍"、"并行跑"、"create team"、"spawn agents"、"组队"时触发。 |
使用 Claude Code 的 Agent Teams 功能创建多 agent 团队，支持并行任务执行和异步 Discord 通知。
TeamCreate → 并行 Task(background, bypassPermissions) → agent 完成后通知 lead → lead 汇总/决策 → lead 通知用户 → shutdown → TeamDelete
TeamCreate → team_name, description
每个 teammate 用 Task tool 启动，必须带以下参数：
team_name: 团队名name: 成员名subagent_type: "general-purpose"mode: "bypassPermissions"（必须，否则权限弹窗卡住）run_in_background: true你是「{名字}」，{team_name} 团队成员，通用型人才。
**日志规则**（必须遵守）：
- 日志文件：~/.claude/teams/{team_name}/logs/{名字}.log
- 启动时创建目录：mkdir -p ~/.claude/teams/{team_name}/logs
- 每次执行操作前后都用 Bash 追加日志：
echo "[$(date '+%H:%M:%S')] {动作描述}" >> ~/.claude/teams/{team_name}/logs/{名字}.log
- 日志内容包括：收到任务、开始执行、执行进度、完成结果、遇到错误
启动后立刻做两件事：
1. mkdir -p ~/.claude/teams/{team_name}/logs
2. echo "[$(date '+%H:%M:%S')] 【{名字}】已上线，待命中" >> ~/.claude/teams/{team_name}/logs/{名字}.log
3. 用 SendMessage 告诉 team-lead 你已就位
然后等待 team-lead 分配任务。
**禁止直接联系用户**：不要使用 send-to-discord.sh，不要直接给用户发消息。所有沟通必须经过 team-lead。
**任务完成后必须执行两步通知**（缺一不可）：
1. 写日志：echo "[$(date '+%H:%M:%S')] 任务完成：{结果摘要}" >> 日志文件
2. 通知 lead：SendMessage 给 team-lead，包含结果摘要和产出物路径
lead 会决定是否需要通知用户。你只对 lead 汇报。
每个 teammate 有独立日志文件，方便实时追踪和 debug：
~/.claude/teams/{team_name}/logs/
├── shunshen.log
├── felix.log
└── chris.log
实时查看：tail -f ~/.claude/teams/{team_name}/logs/*.log
单人查看：tail -f ~/.claude/teams/{team_name}/logs/shunshen.log
核心原则：组员 → Lead → 用户。组员禁止直接联系用户。
通知链：
agent → lead（SendMessage，唯一合法路径）lead → 用户（lead 汇总/过滤后决定是否通知用户）Lead 的职责：
send-to-discord.sh 通知用户（只有 lead 有权使用）SendMessage → type: "message", recipient: "{name}", content: "任务描述"
SendMessage → type: "shutdown_request" → 每个成员
等 shutdown_approved
TeamDelete
rm -rf ~/.claude/teams/{name} 清理