
Multi Agent Architecture
- 37 installs
- 82 repo stars
- Updated August 2, 2026
- aaaaqwq/claude-code-skills
multi-agent-architecture is a Claude Code skill for designing multi-agent systems in OpenClaw, with specialized agents and a smart spawn system that routes tasks to the right agent.
About
multi-agent-architecture is a Claude Code skill for designing multi-agent systems in OpenClaw. It lays out a main agent that dispatches to specialized agents (news, code, research, quick, batch, monitor), each with its own model, system prompt, tools, and channel bindings. It shows how to register agents and implements a smart spawn system that routes tasks by type, complexity, time-sensitivity, and cost. A developer uses it to build a coordinated multi-agent setup. The SKILL.md is written in Chinese.
- Designs multi-agent systems with specialized agents (news, code, research, quick, batch, monitor)
- Defines per-agent model, system prompt, tool, and channel configuration in OpenClaw
- Implements smart spawn that routes tasks to the right agent by type, complexity, and cost
Multi Agent Architecture by the numbers
- 37 all-time installs (skills.sh)
- Ranked #8,484 of 16,556 AI & Agent Building skills by installs in the Skillselion catalog
- Data as of Aug 3, 2026 (Skillselion catalog sync)
multi-agent-architecture capabilities & compatibility
- Capabilities
- multi agent design · agent orchestration · task routing · sub agent spawn
- Use cases
- orchestration · planning
- Pricing
- Free
What multi-agent-architecture says it does
多 Agent 架构设计与智能 Spawn 系统。当需要设计多 Agent 系统、配置专业化 Agent、实现智能任务分发、或优化并发处理能力时使用此技能。
Main Agent 根据任务类型自动选择合适的 Agent
npx skills add https://github.com/aaaaqwq/claude-code-skills --skill multi-agent-architectureAdd your badge
Show developers this skill is listed on Skillselion. Paste this into your README.
| Installs | 37 |
|---|---|
| repo stars | ★ 82 |
| Last updated | August 2, 2026 |
| Repository | aaaaqwq/claude-code-skills ↗ |
What it does
Design and configure a multi-agent system with specialized agents and smart task routing in OpenClaw.
Who is it for?
Designing a multi-agent system with specialized agents and task routing
Skip if: Single-agent workflows that need no delegation
When should I use this skill?
You are designing a multi-agent system, configuring specialized agents, or implementing task dispatch
What you get
A coordinated multi-agent setup where a main agent routes each task to a specialized agent with the right model and tools.
- Multi-agent directory and config structure
- Per-agent model, prompt, and tool configuration
- Smart spawn routing rules and examples
By the numbers
- 6 specialized agent roles (news, code, research, quick, batch, monitor)
- Task routing table across 6 task types
Files
Multi-Agent Architecture - 多 Agent 架构
概述
OpenClaw 支持多 Agent 架构,每个 Agent 可以有不同的:
- 专业领域和 System Prompt
- 模型配置和成本策略
- Channel 绑定和权限
- 工具集和 MCP 配置
架构设计
推荐的 Agent 分工
┌─────────────────────────────────────────────────────────────┐
│ Main Agent (小a) │
│ - 主会话处理 │
│ - 任务分发和协调 │
│ - 复杂决策和规划 │
│ - 模型: opus-4.5 (高质量) │
└─────────────────────────────────────────────────────────────┘
│
┌───────────────────┼───────────────────┐
▼ ▼ ▼
┌─────────────────┐ ┌─────────────────┐ ┌─────────────────┐
│ News Agent │ │ Code Agent │ │ Research Agent │
│ - 新闻抓取 │ │ - 代码生成 │ │ - 深度研究 │
│ - 内容摘要 │ │ - Bug 修复 │ │ - 文档分析 │
│ - 定时推送 │ │ - 代码审查 │ │ - 知识整合 │
│ 模型: sonnet │ │ 模型: codex │ │ 模型: opus │
└─────────────────┘ └─────────────────┘ └─────────────────┘
│ │ │
▼ ▼ ▼
┌─────────────────┐ ┌─────────────────┐ ┌─────────────────┐
│ Quick Agent │ │ Batch Agent │ │ Monitor Agent │
│ - 快速问答 │ │ - 批量处理 │ │ - 系统监控 │
│ - 简单任务 │ │ - 数据处理 │ │ - 健康检查 │
│ - 低延迟响应 │ │ - 文件操作 │ │ - 告警通知 │
│ 模型: flash │ │ 模型: mini │ │ 模型: mini │
└─────────────────┘ └─────────────────┘ └─────────────────┘Agent 配置
1. 创建 Agent 目录结构
~/.openclaw/agents/
├── main/ # 主 Agent (已存在)
│ └── agent/
│ ├── AGENTS.md
│ ├── SOUL.md
│ └── ...
├── news/ # 新闻 Agent
│ └── agent/
│ ├── AGENTS.md
│ └── config.json
├── code/ # 代码 Agent
│ └── agent/
│ ├── AGENTS.md
│ └── config.json
├── research/ # 研究 Agent
│ └── agent/
│ ├── AGENTS.md
│ └── config.json
├── quick/ # 快速响应 Agent
│ └── agent/
│ └── config.json
└── batch/ # 批量处理 Agent
└── agent/
└── config.json2. Agent 配置示例
News Agent (~/.openclaw/agents/news/agent/config.json)
{
"model": {
"primary": "anthropic/claude-sonnet-4-5"
},
"systemPrompt": "你是新闻抓取和摘要专家。专注于:\n1. 从权威来源抓取真实新闻\n2. 生成简洁准确的摘要\n3. 确保每条新闻有原文链接\n4. 按时推送到指定渠道",
"tools": {
"allow": ["web_fetch", "exec", "message"]
}
}Code Agent (~/.openclaw/agents/code/agent/config.json)
{
"model": {
"primary": "openrouter-vip/gpt-5.2-codex"
},
"systemPrompt": "你是代码专家。专注于:\n1. 高质量代码生成\n2. Bug 分析和修复\n3. 代码审查和优化\n4. 技术文档编写",
"tools": {
"allow": ["read", "write", "edit", "exec"]
}
}Quick Agent (~/.openclaw/agents/quick/agent/config.json)
{
"model": {
"primary": "google/gemini-flash-latest"
},
"systemPrompt": "你是快速响应助手。特点:\n1. 简洁直接的回答\n2. 低延迟响应\n3. 处理简单查询\n4. 不需要深度分析的任务"
}3. 在 openclaw.json 中注册 Agent
{
"agents": {
"entries": {
"news": {
"enabled": true,
"allowSpawnFrom": ["main"]
},
"code": {
"enabled": true,
"allowSpawnFrom": ["main"]
},
"research": {
"enabled": true,
"allowSpawnFrom": ["main"]
},
"quick": {
"enabled": true,
"allowSpawnFrom": ["main"]
},
"batch": {
"enabled": true,
"allowSpawnFrom": ["main"]
}
},
"defaults": {
"maxConcurrent": 4,
"subagents": {
"maxConcurrent": 8
}
}
}
}智能 Spawn 系统
任务分类规则
Main Agent 根据任务类型自动选择合适的 Agent:
| 任务类型 | 关键词 | 目标 Agent | 模型 |
|---|---|---|---|
| 新闻抓取 | news, 新闻, 早报, 推送 | news | sonnet |
| 代码任务 | code, 代码, bug, 开发 | code | codex |
| 深度研究 | research, 分析, 调研 | research | opus |
| 快速问答 | 简单, 快速, 查询 | quick | flash |
| 批量处理 | batch, 批量, 文件 | batch | mini |
| 复杂任务 | 保留在 main | main | opus |
智能 Spawn 实现
# 在 AGENTS.md 中添加智能 Spawn 逻辑
## 🧠 智能任务分发
当收到任务时,评估以下因素:
1. **任务复杂度**
- 简单查询 → quick agent
- 中等任务 → 专业 agent
- 复杂任务 → main 处理或 research agent
2. **任务类型**
- 新闻相关 → news agent
- 代码相关 → code agent
- 研究分析 → research agent
- 批量操作 → batch agent
3. **时间敏感度**
- 需要快速响应 → quick agent
- 可以等待 → 专业 agent
4. **资源消耗**
- 高 token 消耗 → 使用便宜模型的 agent
- 需要高质量 → 使用 opus 的 agent
### Spawn 命令示例
新闻任务
sessions_spawn( task="抓取今日科技新闻并推送到 DailyNews 群组", agentId="news", label="news-morning" )
代码任务
sessions_spawn( task="修复 auth.py 中的登录 bug", agentId="code", label="fix-auth-bug" )
研究任务
sessions_spawn( task="深度分析 GPT-5 的技术架构", agentId="research", label="gpt5-analysis" )
快速查询
sessions_spawn( task="查询今天的天气", agentId="quick", label="weather-check" )
## 并发处理
### 配置并发限制
{ "agents": { "defaults": { "maxConcurrent": 4, // 主 agent 最大并发 "subagents": { "maxConcurrent": 8 // 子 agent 最大并发 } } } }
### 并发场景
用户消息 → Main Agent │ ├─→ spawn(news) ──→ 抓取新闻 │ ├─→ spawn(code) ──→ 修复 bug │ └─→ spawn(research) ──→ 深度分析
↓ (并行执行)
所有任务完成后汇报
## Channel 绑定
### 不同 Channel 使用不同 Agent
{ "channels": { "telegram": { "defaultAgent": "main" }, "whatsapp": { "defaultAgent": "main" } }, "agents": { "entries": { "news": { "channels": ["telegram-newsbot"] } } } }
## 监控和管理
### 查看活跃 Session
列出所有 session
openclaw sessions list
查看特定 agent 的 session
openclaw sessions list --agent news
### 查看 Spawn 状态
在代码中
sessions_list(kinds=["spawn"], limit=10)
## 最佳实践
### 1. 任务分发原则
- **简单任务不 spawn** - 直接处理更快
- **耗时任务必 spawn** - 不阻塞主会话
- **相关任务批量 spawn** - 提高效率
### 2. 模型选择原则
- **质量优先** → opus
- **速度优先** → flash
- **代码任务** → codex
- **成本优先** → mini
### 3. 错误处理
spawn 时设置超时
sessions_spawn( task="...", agentId="code", runTimeoutSeconds=300, # 5分钟超时 cleanup="keep" # 保留 session 用于调试 )
## 相关资源
- [OpenClaw Agents 文档](https://docs.openclaw.ai/agents)
- [Sessions Spawn 文档](https://docs.openclaw.ai/tools/sessions)
- [模型配置指南](https://docs.openclaw.ai/models)
---
*由小a设计 - 实现真正的多 Agent 协作*
Related skills
FAQ
What agents does the architecture propose?
A main agent that dispatches to news, code, research, quick, batch, and monitor agents, each with its own model and tools.
How are tasks assigned to agents?
A smart spawn system routes tasks by type, complexity, time-sensitivity, and resource cost.