synapse-brain
Synapse Brain — OpenClaw 持久调度 Agent。 基于 Managed Agents 架构,提供跨 Session 的任务管理、子代理调度、 状态持久化和知识互操作能力。是 synapse-code 和 synapse-wiki 的调度核心。 当用户提到任务调度、跨会话管理、多 Agent 协作、长期项目跟踪时使用此技能。
Install via CLI (Recommended)
clawhub install openclaw/skills/skills/ankechenlab-node/synapse-brainSynapse Brain Skill
Synapse Brain = 持久调度 + 跨会话状态 + 子代理编排
核心理念:AI Agent 应该是持久的"大脑",而非每次从零开始的无状态会话。
| 传统 Agent | Synapse Brain | |
|---|---|---|
| 会话 | 每次从零开始 | 状态持久化,醒来即恢复 |
| 调度 | 手动调用子代理 | 自动路由、分配、追踪 |
| 记忆 | 随会话消失 | state.json 持久化 |
| 协作 | 独立工作 | 调度 code/wiki 等 Skills |
🚦 快速决策:我该用什么?
你想做什么?
│
├─ 启动/恢复项目任务 → init(初始化 session)
│ └─ 例:"继续做用户系统"、"开始新项目 X"
│
├─ 分配子任务给子代理 → dispatch(调度执行)
│ └─ 例:"并行开发登录和注册"、"让 dev 实现导出"
│
├─ 查询任务状态 → status(状态查询)
│ └─ 例:"现在进度如何?"、"昨天那个任务完成了没"
│
└─ 保存/恢复会话 → save/restore(状态管理)
└─ 例:"保存当前进度"、"恢复上次会话"
📋 命令速查卡片
| 命令 | 用途 | 示例 |
|---|---|---|
/synapse-brain init | 初始化/恢复项目 session | /synapse-brain init my-project "用户系统" |
/synapse-brain dispatch | 分发任务到子代理 | /synapse-brain dispatch "并行开发 A 和 B" |
/synapse-brain status | 查询任务/子代理状态 | /synapse-brain status my-project |
/synapse-brain save | 保存当前会话状态 | /synapse-brain save |
/synapse-brain restore | 恢复上次会话 | /synapse-brain restore |
Brain Agent 架构
用户请求
↓
┌─────────────────────────────┐
│ Synapse Brain (持久调度) │
│ │
│ 1. Session Restore │ ← state.json 恢复上下文
│ 2. Intent Classification │ ← 识别意图类型
│ 3. Task Routing │ ← 路由到合适 Skill/Agent
│ 4. Subagent Dispatch │ ← 调用 OpenClaw 子代理
│ 5. Result Aggregation │ ← 汇总结果
│ 6. State Persist │ ← 保存状态到 state.json
└─────────────┬───────────────┘
↓
┌─────────┼─────────┐
↓ ↓ ↓
synapse- synapse- 专业子代理
code wiki (dev/qa/arch)
Session 持久化
Brain 通过 state.json 维护跨会话状态:
{
"session_id": "sess_20260410_001",
"project": "my-project",
"created_at": "2026-04-10T12:00:00Z",
"updated_at": "2026-04-10T14:30:00Z",
"tasks": [
{
"id": "task-001",
"title": "用户登录功能",
"status": "completed",
"skill": "synapse-code",
"mode": "lite",
"agents_used": ["req-analyst", "developer", "qa-engineer"]
},
{
"id": "task-002",
"title": "用户注册功能",
"status": "in_progress",
"skill": "synapse-code",
"mode": "lite",
"stage": "DEV"
}
],
"subagents": {
"active": 1,
"completed": 3,
"failed": 0
},
"knowledge": {
"wiki_initialized": true,
"wiki_root": "~/my-project/wiki",
"last_ingest": "2026-04-10T13:00:00Z"
}
}
子代理调度策略
Brain 根据任务特征自动选择调度策略:
| 策略 | 触发条件 | 行为 |
|---|---|---|
| sequential | 任务有依赖关系 | 依次执行子代理 |
| parallel | 任务独立可并行 | 同时派发最多 8 个子代理 |
| pipeline | 标准开发流程 | REQ→ARCH→DEV→INT→QA→DEPLOY |
| standalone | 简单任务 | Brain 直接完成,不调子代理 |
与 synapse-code 互操作
# Brain 调度 code 执行开发任务
/synapse-brain dispatch "实现用户登录" --skill synapse-code --mode lite
# 等价于
/synapse-code run my-project "实现用户登录" --mode lite
Metadata
Not sure this is the right skill?
Describe what you want to build — we'll match you to the best skill from 16,000+ options.
Find the right skillPaste this into your clawhub.json to enable this plugin.
{
"plugins": {
"official-ankechenlab-node-synapse-brain": {
"enabled": true,
"auto_update": true
}
}
}Tags
Related Skills
calling-agent-squad
Activate a multi-agent team (the Squad) to manage complex projects, business tasks, or development workflows. The squad includes a Manager, Architect, Coder, Reviewer, and Observer. Use when the user wants to "call a squad", "start a project", or "deploy squad" with specialized roles and quality control loops.
agent-health-diagnostics
Diagnose and fix the 4 most common OpenClaw agent failures — heartbeat spam, API rate limit cascades, channel death loops, and memory/embedding errors. Battle-tested across a 6-agent multi-host deployment.
synapse-code
Synapse Code — 智能代码开发工作流引擎。 一体化完成项目初始化、代码交付、知识沉淀和影响分析。 内建代码图谱引擎,越用越懂你的项目。 当用户提到开发、实现功能、运行 pipeline、记录知识、检查影响范围时使用此技能。
multi-agent-status
Cross-agent health monitoring for multi-host OpenClaw deployments. Each agent pushes structured status reports (JSON) to a central location. A PM/monitoring agent reads them and alerts on failures. Works across Windows, Linux, and mixed environments.
memory-optimizer-base
多Agent记忆管理系统 - 开放协作的知识库解决方案 支持私有+公共双层记忆空间,自动生成每日总结,跨Agent知识检索