En

Agent Comm Hub

Multi-Agent Real-Time Communication and Task Orchestration Hub

开发流程 协调多智能体编排工作流程分配团队任务 通用 ★ 2.4k Updated 2026-08-02

Install & Use

Copy this prompt and send it to your AI assistant (Claude / Cursor / TRAE / Codex / WorkBuddy etc.) to auto-install:

Help me install this AI Skill: Agent Comm Hub.
It is used for: Multi-Agent Real-Time Communication and Task Orchestration Hub
Full Skill content: https://321skill.com/skills/agent-comm-hub-x-8/raw/index.md
Read that page and install it.

The prompt includes a link to the full Skill content. You can also view the full content.

This Skill addresses the pain point of lacking unified communication and context sharing among multiple AI Agents. In practical development, developers often need multiple independent AI Agents to collaborate, but frequently encounter issues such as high message latency, unsynchronized task states, and fragmented memory, which prevent agents from working together efficiently.

Usage is straightforward: simply start the Agent Communication Hub as an MCP server, then connect your agents via stdio or HTTP-SSE protocols. You can then use natural language commands to register agents, send peer-to-peer messages, assign tasks, store shared memories, and leverage 58 MCP tools to view online status, manage files, perform task orchestration, and more. The entire process has a latency under 50ms, with messages persisted in SQLite.

It is well-suited for teams or individuals needing to coordinate multiple AI agents, especially developers already building agents using the MCP protocol who require task orchestration and shared memory capabilities. Whether building customer service bot clusters, automating workflows, or researching multi-agent collaboration models, users can benefit from it.

It is recommended to first register an agent using an invite_code during initial setup and maintain a heartbeat every 30 seconds to stay online. Note that it relies on local SQLite storage, making it suitable for single-machine or small-scale deployments, not for large-scale distributed scenarios across regions.

Key Features

Unlike typical MCP servers (e.g., single-tool providers), the Agent Comm Hub comes with a built-in message queue, task state machine, shared memory, RBAC permission management, and a web admin panel. Comparable tools like `mcp-server-sse` only provide basic SSE channels, lacking task orchestration and persistence capabilities.

Limitations

The current version is based on Node.js and SQLite WAL mode, designed for single-machine deployment. It does not support horizontal scaling via clustering and requires agents to actively maintain a heartbeat (agents go offline automatically after 90 seconds without a heartbeat).

FAQ

How do I register a new Agent?

Call the `register_agent` tool, providing an `invite_code` and an Agent name. The system will assign a unique `agent_id`.

What connection methods does the Hub support?

It supports two modes: MCP stdio (for local processes) and HTTP-SSE (for remote connections). Agents communicate via JSON-RPC.

Installation guide for AI assistants

If your AI coding assistant (Claude Code, Cursor, TRAE etc.) can see this page, send it this message to auto-install:

Visit https://321skill.com/skills/agent-comm-hub-x-8/raw/index.md to read the original Skill definition (Markdown format) for Agent Comm Hub, and install it according to the instructions.