agent-bridge
Bridge AI coding agents across machines -- works with Claude Code, Codex, OpenClaw, Gemini CLI, Aider, and any CLI agent.
⚠️ Tested end-to-end with Claude Code only (as of v2.3.3, 2026-04-14). The Codex / Gemini CLI / OpenClaw / Aider integrations are designed to work via MCP config but haven't been exercised by the author yet. Use on those harnesses at your own risk and please report back.
Website | GitHub
What is agent-bridge?
agent-bridge lets AI coding agents on different machines talk to each other and run commands on each other's machines over SSH. It is:
- Peer-to-peer -- no central server, no cloud, just direct SSH between machines
- Agent-agnostic -- works with Claude Code, Codex, Gemini CLI, OpenClaw, Aider, or any CLI agent
- Zero dependencies -- just bash, ssh, and ssh-keygen (built into every Mac and Linux)
- Real-time -- messages are pushed into running Claude Code and OpenClaw sessions automatically (channel plugins), or polled by other harnesses
Architecture overview
agent-bridge architecture
MACHINE A (e.g. Mac Mini) MACHINE B (e.g. MacBook Pro)
┌─────────────────────────────────┐ ┌─────────────────────────────────┐
│ │ │ │
│ AI Agent (Claude Code, etc.) │ │ AI Agent (Claude Code, etc.) │
│ ┌───────────────────────────┐ │ │ ┌───────────────────────────┐ │
│ │ MCP Server / Channel │ │ SSH │ │ MCP Server / Channel │ │
│ │ ┌───────────────────────┐ │ │◄───────►│ │ ┌───────────────────────┐ │ │
│ │ │ bridge_send_message │ │ │messages │ │ │ bridge_send_message │ │ │
│ │ │ bridge_run_command │ │ │ │ │ │ bridge_run_command │ │ │
│ │ │ bridge_status │ │ │ │ │ │ bridge_status │ │ │
│ │ │ ... │ │ │ │ │ │ ... │ │ │
│ │ └───────────────────────┘ │ │ │ │ └───────────────────────┘ │ │
│ │ File watcher (inbox) │ │ │ │ File watcher (inbox) │ │
│ └───────────────────────────┘ │ │ └───────────────────────────┘ │
│ │ │ │
│ agent-bridge CLI │ │ agent-bridge CLI │
│ ~/.agent-bridge/ │ │ ~/.agent-bridge/ │
│ config, keys/, inbox/ │ │ config, keys/, inbox/ │
└─────────────────────────────────┘ └─────────────────────────────────┘
Both machines are PEERS -- either can run commands on the other.
No fixed controller or target.
Compatibility
agent-bridge works with any AI coding agent that can run shell commands or MCP tools. It ships with skill files for the CLI and an MCP server for real-time messaging:
| Agent Harness | Skill File | MCP Messaging | Delivery |
|---|
| Claude Code | skills/bridge/skill.md | Yes (channel plugin) | Push -- messages arrive automatically |
| OpenClaw | skills/openclaw/SKILL.md | Yes (MCP server) + channel plugin | Push (with plugin/daemon) or polling fallback |
| Codex CLI (OpenAI) | AGENTS.md | Yes (MCP server) | Polling via bridge_receive_messages |
| Gemini CLI | GEMINI.md | Yes (MCP server) | Polling via bridge_receive_messages |
| Aider / others | INSTRUCTIONS.md | Yes (MCP server) | Polling via bridge_receive_messages |
Each skill file teaches the respective agent how to use agent-bridge: listing machines, running commands, pairing from photos, and delegating work to remote agents. The MCP server adds real-time messaging between running agent sessions.
Quick start
Paste this into your AI agent on each computer you want to bridge:
Read the README at https://github.com/EthanSK/agent-bridge and follow the setup instructions
for this computer. Install agent-bridge, run the setup command, and install the skill for this
agent. Do everything automatically -- don't ask me questions.
