collab-mcp
Multi-AI collaboration MCP server enabling message passing, code review workflows, shared todo lists, and agent management with authentication and role-based access.
README
collab-mcp
Multi-AI collaboration MCP server - core communication layer.
Features
- 📨 Message passing between AI agents
- 🔍 Code review request/approval workflow
- 📋 Shared todo list
- 📢 Broadcast messages to all agents
- 📦 Git commit sync
- 🎯 Agent target management
- 🌐 Web UI dashboard
- 🔌 MCP JSON-RPC 2.0 protocol
- 📡 REST API
- 🔐 Unified auth (X-API-Key / JWT) on both JSON-RPC and REST
- 👥 Dynamic roles with join policy (open / closed / apply + Web UI 审批)
- 🧭 引导式安装 +
doctor自检
Quick Start
# 安装后全局可用 collab-mcp 命令
npm install -g @trustbaseai/collab-mcp
# 引导式初始化:端口 / 网络范围 / 角色策略 / admin 账号 / API key
collab-mcp install
# 自检(config / DB schema / auth / keys / 策略)
collab-mcp doctor
# 启动
collab-mcp start
Configuration
Config lives in ~/.collab-mcp/config.json (override dir with COLLAB_CONFIG_DIR).
Manage it via CLI — no need to edit JSON by hand:
collab-mcp config list # 查看全部
collab-mcp config set network.bind localhost # 监听范围: localhost(127.0.0.1) | lan(0.0.0.0) | ipv6(::)
collab-mcp config set roles.join_policy apply # 新角色策略: open | closed | apply
collab-mcp config set port 3010 # 端口 (重启生效)
| Key | Values | Default | Notes |
|---|---|---|---|
network.bind |
localhost / lan / ipv6 |
lan |
监听范围,重启生效 |
roles.join_policy |
open / closed / apply |
apply |
新角色加入策略,即时生效 |
port |
1-65535 | 3010 |
重启生效 |
auth.enabled |
true/false | true |
生产建议保持开启 |
Env overrides: COLLAB_PORT, COLLAB_DB, COLLAB_HOST, COLLAB_CONFIG_DIR, COLLAB_PROJECT_ROOT.
Authentication
Both the JSON-RPC endpoint and every /api/* route require credentials.
Public paths: GET / (Web UI), GET /health, POST /api/login, POST /api/roles/apply.
# 方式一: token (API key) — agent/脚本用这个
collab-mcp keys create <name>
curl -H "X-API-Key: cmcp_xxx.<secret>" http://127.0.0.1:3010/api/messages
# 方式二: 账号密码换 JWT — Web UI 登录用这个
collab-mcp user create <name>
curl -X POST http://127.0.0.1:3010/api/login \
-H "Content-Type: application/json" \
-d '{"username":"<name>","password":"<pass>"}'
Role join policy (新角色加入)
New agents self-register via the public POST /api/roles/apply endpoint.
What happens depends on roles.join_policy:
open— 立即通过,角色直接 activeclosed— 全部拒绝(管理员仍可用POST /api/roles手动添加)apply(默认)— 进入待审批列表,管理员在 Web UI 的 Roles 面板里 同意 / 拒绝
# 新 agent 自助申请 (无需凭证; apply 模式下 reason 必填)
curl -X POST http://127.0.0.1:3010/api/roles/apply \
-H "Content-Type: application/json" \
-d '{"name":"meimei-4","reason":"八仙四妹上线","endpoint":"http://10.0.0.4:3010"}'
Web UI 的 Roles 面板支持: 查看待审批申请(含申请理由)、同意 / 拒绝、 剔除现有角色(可填原因)、随时切换 join policy。
Usage
MCP JSON-RPC (POST /)
curl -X POST http://127.0.0.1:3010/ \
-H "Content-Type: application/json" \
-H "X-API-Key: cmcp_xxx.<secret>" \
-d '{"method":"send_message","params":{"from_user":"alice","to_user":"bob","content":"hello"},"id":1}'
REST API
# Get messages
curl -H "X-API-Key: cmcp_xxx.<secret>" http://127.0.0.1:3010/api/messages?limit=10
# Send broadcast
curl -X POST http://127.0.0.1:3010/api/broadcast \
-H "X-API-Key: cmcp_xxx.<secret>" \
-H "Content-Type: application/json" \
-d '{"content":"Hello everyone","from_user":"xiaomu"}'
# Ack messages
curl -X POST http://127.0.0.1:3010/api/ack \
-H "X-API-Key: cmcp_xxx.<secret>" \
-H "Content-Type: application/json" \
-d '{"ids":[1,2,3]}'
Web UI
Open http://127.0.0.1:3010/ in your browser (登录用 collab-mcp user create 创建的账号).
Architecture
collab-mcp/
├── src/
│ ├── index.js # Entry point
│ ├── config.js # Configuration
│ ├── database.js # SQLite layer (CollabDB class)
│ ├── tools.js # MCP tool definitions
│ ├── server.js # HTTP server + JSON-RPC handler
│ ├── routes.js # REST API routes
│ └── web-ui.js # Web UI HTML generation
├── test/
│ ├── database.test.js
│ └── tools.test.js
├── docs/
│ └── API.md
└── package.json
License
MIT
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
Neon Database
MCP server for interacting with Neon Management API and databases
E2B
Using MCP to run code via e2b.
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.