← 返回
AI智能 中文

OpenClaw Swarm

Enable advanced subagent orchestration with context sharing, state management, parallel tasks, and completion notifications using OpenClaw Swarm features.
借助 OpenClaw Swarm 功能实现高级子智能体编排,支持上下文共享、状态管理、并行任务与完成通知。
heldinhow
AI智能 clawhub v1.0.2 1 版本 99897.5 Key: 无需
★ 0
Stars
📥 975
下载
💾 51
安装
1
版本
#latest

概述

openclaw-swarm

Use OpenClaw Swarm features for advanced subagent orchestration.

About

This skill provides access to OpenClaw Swarm - a fork of OpenClaw with enhanced subagent orchestration:

  • Fork: https://github.com/Heldinhow/openclaw-swarm
  • Docs: SWARM.md

When to Use

Use this skill when:

  • Spawning subagents with context sharing
  • Coordinating multiple subagents
  • Sharing state between subagents
  • Running parallel tasks

Tools Available

1. sessions_spawn with contextSharing

Share parent session context with subagents:

{
  "sessions_spawn": {
    "label": "my-task",
    "task": "Do something",
    "contextSharing": "recent"
  }
}

Values:

  • none - No context
  • summary - Compressed summary
  • recent - Last messages
  • full - Complete history

2. context_store

Share data between subagents:

// Write
{ "context_store": { "action": "set", "namespace": "project", "key": "data", "value": {...} } }

// Read
{ "context_store": { "action": "get", "namespace": "project", "key": "data" } }

Actions: get, set, delete, list, subscribe, broadcast

3. context_publish

Notify when subagent completes:

{ "context_publish": {
    "action": "publish",
    "eventType": "task_complete",
    "target": "orchestrator",
    "data": { "result": "..." }
  }
}

4. parallel_spawn

Run multiple subagents in parallel:

{ "parallel_spawn": {
    "tasks": [
      { "label": "task1", "task": "Do this" },
      { "label": "task2", "task": "Do that" }
    ],
    "wait": "all"
  }
}

Wait strategies:

  • all - Wait for all
  • any - Return on first, others continue
  • race - Return on first

Patterns

Parallel Research

{
  "parallel_spawn": {
    "tasks": [
      { "label": "web-search", "task": "Search X" },
      { "label": "docs-search", "task": "Find docs about X" }
    ],
    "wait": "all"
  }
}

Chain Workflow

  1. First subagent writes to context_store
  2. Second subagent reads from context_store
  3. Both notify via context_publish

Context Pipeline

{
  "sessions_spawn": {
    "label": "processor",
    "task": "Process data",
    "contextSharing": "recent"
  }
}

Auto-Announce

Subagents automatically announce completion:

✅ Sub-agent completed: label
   task: ...
   result: ...
   runtime: Xs

No polling needed!

版本历史

共 1 个版本

  • v1.0.2 当前
    2026-03-29 02:44 安全 安全

安全检测

腾讯云安全 (Keen)

安全,无风险
查看报告

腾讯云安全 (Sanbu)

安全,无风险
查看报告

🔗 相关推荐

ai-intelligence

ontology

oswalpalash
类型化知识图谱,用于结构化智能体记忆与可组合技能。支持创建/查询实体(人员、项目、任务、事件、文档)及关联...
★ 714 📥 244,100
ai-intelligence

self-improving agent

pskoett
捕获经验教训、错误和纠正,以实现持续改进。使用时机:(1)命令或操作意外失败;(2)用户纠正……
★ 4,061 📥 799,449
ai-intelligence

Self-Improving + Proactive Agent

ivangdavila
自我反思+自我批评+自我学习+自组织记忆。智能体评估自身工作、发现错误并持续改进。
★ 1,362 📥 318,964