- Home
- Skills
- Agents & Orchestration
- agent msg bus
agent msg bus
A local multi-agent message bus for persistent agent communication, broadcasting, and background standby loops.
$10
Works with the AI Agents you already use
agent msg bus
Example session with this skill installed
Join the message platform as msgbus-refactor and listen for messages.
- Read your context and instructions
- Compiled the agent msg bus
I have initialized the message broker and joined the platform.
I'm online as msgbus-refactor; others can reach me at @msgbus-refactor. I am now entering the listen loop to wait for incoming messages.
Session Details
| Attribute | Value |
|---|---|
| Agent Name | @msgbus-refactor |
| Status | Online / Listening |
| Broker Path | node "agent-msg-bus/msg.js" |
| Timeout | 540s (9 minutes) |
| Human User | @user |
Next steps
- To see other agents online, run
node "agent-msg-bus/msg.js" who. - To send a message, use
node "agent-msg-bus/msg.js" send @peer --as msgbus-refactor "text". - Tell me to "stop listening" whenever you want to terminate the standby loop.
Connects securely to your tools. The creator never sees your data.
About this skill
Multi-agent systems often fail because agents cannot communicate asynchronously or stay online to wait for peers. This skill provides a lightweight, local message bus that allows agents to join a platform, exchange messages, and maintain a listening loop for background collaboration.
What it does
- Persistent Sessions allows an agent to stay online and listen for incoming messages from peers or the human user.
- Targeted Messaging enables sending messages to specific @named agents or broadcasting to @all participants.
- Broker Management includes a built-in broker with zero dependencies to handle message queuing and delivery.
- Asynchronous Waiting supports a non-blocking recv loop with configurable timeouts to standby for new tasks.
- Peer Discovery provides a directory of currently active agents via the who command.
How it works
- Join the bus by initializing the broker and picking a unique two-word shortname for the current session.
- Exchange data using send and recv commands, including a combined send-and-wait mode for efficient round-trips.
- Listen loop keeps the agent active for up to 18 minutes of idle time, processing incoming requests as they arrive.
- Broadcast updates to notify all other online agents of state changes or task completions.
Frameworks & tools
This tool uses a Node.js runtime to execute msg.js and broker.js. It operates entirely through shell commands and local file-based communication, requiring no external APIs or heavy dependencies.
Why this beats prompting it yourself
Prompting an agent to "talk to another agent" usually results in a one-off call that loses context or fails when the other agent is busy. This skill provides the actual networking layer, ensuring messages are queued and agents know exactly how to listen and reply without getting stuck in a loop.
Use cases
- Parallel Research where one agent performs deep analysis in the background while another handles user queries.
- Coordinated Refactoring across multiple agents specializing in different parts of a codebase.
- Standing Standby for agents that need to wait for human approval or long-running CI/CD triggers.
Known limitations
Requires Node.js installed on the host. Not intended for one-shot stateless delegation. Name collisions require manual retry logic.
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Agents & Orchestration.
Requires Node.js. Not for one-shot stateless delegation. Agent names must be unique alphanumerics under 20 characters. Listening loop capped at ~18 minutes of silence.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before buying a skill.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 25 days ago
- Software
Node.js
Permissions
Allowed Hosts