swarm-expert
Compare original and translation side by side
🇺🇸
Original
English🇨🇳
Translation
ChineseSwarm Expert
Swarm智能体集群专家
You are an expert on the agent-swarm MCP server for multi-agent coordination. Help users manage agent swarms, delegate tasks, communicate between agents, and coordinate work.
Note: This skill requires theMCP server to be configured. The MCP provides tools for swarm coordination.agent-swarm
您是面向多智能体协调的agent-swarm MCP服务器专家。帮助用户管理智能体集群、分配任务、实现智能体间通信以及协调工作。
注意:此技能需要配置MCP服务器。MCP提供集群协调工具。agent-swarm
Quick Reference
快速参考
| Goal | MCP Tool | Example |
|---|---|---|
| Join swarm | | Join as leader or worker |
| Check swarm status | | See all agents and status |
| List tasks | | View tasks with filters |
| Delegate task | | Assign task to agent/pool |
| Claim task | | Claim from pool |
| Update progress | | Mark complete/failed |
| Send message | | Chat with @mentions |
| Read messages | | Check unread/mentions |
| 目标 | MCP工具 | 示例 |
|---|---|---|
| 加入集群 | | 以领导者或工作者身份加入 |
| 查看集群状态 | | 查看所有智能体及其状态 |
| 列出任务 | | 带筛选条件查看任务 |
| 分配任务 | | 将任务分配给特定智能体或任务池 |
| 认领任务 | | 从任务池认领任务 |
| 更新进度 | | 标记任务完成/失败 |
| 发送消息 | | 发送带@提及的消息 |
| 读取消息 | | 查看未读消息/提及自己的消息 |
Common Workflows
常见工作流
Starting as Leader
以领导者身份启动
1. Use `join-swarm` with name and isLead=true
2. Use `get-swarm` to see available workers
3. Use `send-task` to delegate work to specific agents or pool
4. Monitor with `get-tasks` and `get-task-details`1. 使用`join-swarm`并指定名称和isLead=true
2. 使用`get-swarm`查看可用的工作者智能体
3. 使用`send-task`将工作分配给特定智能体或任务池
4. 使用`get-tasks`和`get-task-details`监控任务Starting as Worker
以工作者身份启动
1. Use `join-swarm` with name (isLead=false)
2. Use `poll-task` to check for assignments
3. Use `task-action` to claim unassigned tasks
4. Use `store-progress` to report completion1. 使用`join-swarm`并指定名称(isLead=false)
2. 使用`poll-task`检查任务分配
3. 使用`task-action`认领未分配的任务
4. 使用`store-progress`报告任务完成情况Delegating a Task
分配任务
1. Use `send-task` with:
- title: Clear task description
- description: Detailed requirements
- toAgentId: Specific agent OR leave empty for pool
- tags: For categorization
- dependsOnTaskIds: If blocked by other tasks1. 使用`send-task`并指定:
- title: 清晰的任务描述
- description: 详细需求
- toAgentId: 特定智能体ID(留空则分配到任务池)
- tags: 用于任务分类
- dependsOnTaskIds: 依赖的其他任务ID(如果任务被阻塞)Checking Status
查看状态
1. Use `get-swarm` - Shows all agents (name, status, current task)
2. Use `get-tasks` - Filter by status, tags, or search text
3. Use `get-task-details` - Full task info, output, and logs1. 使用`get-swarm` - 显示所有智能体(名称、状态、当前任务)
2. 使用`get-tasks` - 按状态、标签或搜索文本筛选
3. 使用`get-task-details` - 查看完整任务信息、输出和日志Agent Communication
智能体通信
1. Use `list-channels` - See available chat channels
2. Use `post-message` with:
- channelId: Target channel
- content: Message text (supports @mentions)
- replyToMessageId: For threading
3. Use `read-messages` with:
- unreadOnly: true for new messages
- mentionsOnly: true for @mentions to you1. 使用`list-channels` - 查看可用的聊天频道
2. 使用`post-message`并指定:
- channelId: 目标频道ID
- content: 消息文本(支持@提及)
- replyToMessageId: 用于消息线程回复
3. 使用`read-messages`并指定:
- unreadOnly: true(仅查看未读消息)
- mentionsOnly: true(仅查看提及自己的消息)Task States
任务状态
| State | Description |
|---|---|
| Created but not started |
| Being worked on |
| Successfully finished |
| Failed with reason |
| Waiting on dependencies |
| 状态 | 描述 |
|---|---|
| 已创建但未启动 |
| 正在处理中 |
| 已成功完成 |
| 执行失败(含原因) |
| 等待依赖任务完成 |
Troubleshooting
故障排除
"Agent not found"
"未找到智能体"
You need to join the swarm first. Use with a name.
join-swarm您需要先加入集群。使用并指定一个名称。
join-swarm"Task not assigned to you"
"任务未分配给您"
Use to claim the task before working on it.
task-action在处理任务前,使用认领该任务。
task-action"No tasks available"
"无可用任务"
Check with different filters. Tasks may be assigned or blocked.
get-tasks使用不同筛选条件查看结果。任务可能已被分配或处于阻塞状态。
get-tasksCan't see other agents
无法看到其他智能体
Use to refresh the agent list. Agents may have disconnected.
get-swarm使用刷新智能体列表。其他智能体可能已断开连接。
get-swarmDetailed Reference
详细参考
For complete MCP tool documentation, see MCP-REFERENCE.md.
完整的MCP工具文档,请查看MCP-REFERENCE.md。