Setup & Installation
Or with the ClawHub CLI, for registry-managed skill folders outside a full OpenClaw workspace:
Version History
Stable agent IDs across restarts, hostname trust migration, dynamic address tracking
What This Skill Does
UDP Messenger lets OpenClaw agents on the same local network discover each other, send messages, and coordinate in real time. It uses UDP sockets with a configurable trust model, so unknown peers require explicit user approval before messages are delivered. An optional webhook hook token enables automatic agent wake-up when a trusted peer sends a message.
Runs entirely on the local network without any cloud service, API key, or external account.
When to use it
- Coordinating tasks between two AI agents on different machines on a home or office LAN
- Discovering which agents are currently running on the local network
- Approving a new peer agent before allowing it to communicate with your agent
- Reviewing the full message history between agents after a multi-agent workflow
- Limiting how many automated message exchanges occur per peer per hour
Example Workflow
Here's how your AI assistant might use this skill in practice.
User asks: Check if any agents are online and ask the one at 192.168.1.42 for the current build status
- 1Call udp_discover to broadcast a ping and list responding agents on the LAN
- 2Call udp_add_peer with 192.168.1.42 to register the target if not yet known
- 3Present the peer to the user and call udp_approve_peer after confirmation
- 4Call udp_send with the target address and the status request message
- 5Call udp_receive to check for a reply and show it to the user
Build status reply received from agent at 192.168.1.42 and displayed to the user