Agent Comm Hub
A local multi-agent communication hub with 58 MCP tools and a web management panel.
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: A local multi-agent communication hub with 58 MCP tools and a web management panel. Full Skill content: https://321skill.com/skills/agent-comm-hub-x-4/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 lack of unified communication and context sharing among multiple AI Agents. In practical development, independent AI Agents often operate in silos, leading to message loss, difficult task orchestration, and unshared memory, which hampers collaboration efficiency. The Agent Comm Hub provides real-time bidirectional communication, task orchestration, shared memory, and an evolution engine via the MCP stdio/HTTP-SSE protocol, enabling Agents to collaborate like a team.
Usage is straightforward: simply start the Hub server, then register Agents and maintain their online status by sending heartbeats via MCP tools. After that, you can use natural language instructions to make Agents send messages, assign tasks, upload files, and store memories. For example: call send_message for Agent A to send a message to Agent B, or call assign_task to create and assign a task. The Hub automatically handles deduplication, persistence, and real-time push notifications.
It is ideal for teams or individuals requiring multi-agent collaboration, especially developers who already have multiple independent AI Agents (e.g., Hermes, WorkBuddy) and want them to work together. Agent developers, backend engineers, and DevOps professionals can all benefit, such as in scenarios like office automation, intelligent customer service, and multi-agent research, where it enables rapid setup of communication infrastructure.
Before first use, it is recommended to prepare an invite_code and ensure the Hub server runs continuously. Note that it primarily relies on the MCP protocol, supporting only stdio and HTTP-SSE modes, and is not suitable for large-scale distributed deployment. Additionally, heartbeats (every 30 seconds) must be managed properly, or Agents will be automatically marked offline.
Key Features
Unlike traditional message queues (e.g., RabbitMQ, Redis Pub/Sub), Agent Comm Hub is designed specifically for AI Agents, featuring built-in shared memory storage, a task orchestration state machine, and an evolution engine. It exposes 58 MCP tools that can be directly invoked by Agents, eliminating the need for additional development of adaptation layers.
Limitations
The current version (v3.0.22) only supports MCP stdio and HTTP-SSE protocols, requires local Hub service operation, and does not support cross-network distributed clusters. It relies on SQLite WAL mode, with a single-machine performance cap of approximately 50ms latency, making it unsuitable for high-throughput scenarios.
FAQ
How do I enable multiple Agents to communicate with each other?
After starting the Hub, each Agent registers and sends heartbeats via MCP tools. Then, simply call send_message for point-to-point real-time communication; messages are automatically persisted and pushed.
What dependencies need to be installed?
A Node.js environment (v18+ recommended) is required. Install the agent-comm-hub package globally via npm, then run the hub service. No additional databases or middleware are needed.
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-4/raw/index.md to read the original Skill definition (Markdown format) for Agent Comm Hub, and install it according to the instructions.
Raw Markdown URL for AI: /skills/agent-comm-hub-x-4/raw/index.md