peer-messaging

Facilitate inter-process communication between AI coding agents via a local MCP broker.

1|Updated May 21, 2026
One-click install
npx skills add https://github.com/hiddink-ai/hiddink-harness --skill peer-messaging-hiddink-ai
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: peer-messaging
Source: https://github.com/hiddink-ai/hiddink-harness/tree/main/templates/skills/peer-messaging
Command: npx skills add https://github.com/hiddink-ai/hiddink-harness --skill peer-messaging-hiddink-ai

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires claude-peers-mcp.

What problem does it solve?

This skill resolves the isolation between multiple AI coding instances, allowing them to communicate and synchronize tasks across different terminal sessions or projects.

Core Features & Use Cases

  • Cross-Session Coordination: Enables real-time messaging between independent Claude Code instances via a local broker.
  • Task Synchronization: Allows agents to broadcast task summaries and receive signals to manage dependencies between concurrent projects.
  • Use Case: When working on a multi-project architecture, use this skill to have an agent in one terminal notify an agent in another terminal that a shared dependency has been updated.

Quick Start

Use the peer-messaging skill to list all active Claude instances and send a status update to the peer working on the dependent project.

Frequently Asked Questions about peer-messaging

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I coordinate tasks between multiple AI coding agents in different terminal sessions?

To synchronize tasks across parallel AI agent workflows, this Skill uses a local broker to broadcast task summaries and send dependency signals. Agents in different terminal sessions receive updates to manage concurrent project dependencies.

How does inter-process communication work for independent Claude Code instances?

Inter-process communication for independent Claude Code instances works via a local MCP broker. The broker manages message routing and state persistence through a local SQLite database, allowing separate sessions to exchange real-time coordination signals.

Do I need to install an MCP broker before setting up cross-session agent messaging?

Yes, setting up cross-session agent messaging requires the claude-peers-mcp broker. This local dependency manages the message routing and SQLite state persistence necessary for independent AI coding instances to discover and communicate with peers.

Can I use local SQLite databases to synchronize state between concurrent AI agent projects?

Yes, you can synchronize state between concurrent AI agent projects using a local SQLite database. The required MCP broker uses SQLite to persist messages and state, enabling agents to recover task context and manage dependencies across sessions.

What are the limitations of using a local MCP broker for multi-agent coordination?

The limitation of using a local MCP broker for multi-agent coordination is that messaging is restricted to independent sessions on the same local machine. It does not support remote agent communication or distributed cloud environments, focusing strictly on local inter-process workflows.