LaunchKit · 2026
Back to Skills

agent-network

Multi-Agent group chat collaboration system inspired by DingTalk/Lark.

3
1.2k downloads
by @howtimeschange

Setup & Installation

openclaw skills install @howtimeschange/agent-network

Or with the ClawHub CLI, for registry-managed skill folders outside a full OpenClaw workspace:

npx clawhub install agent-network

Version History

v1.1.0Feb 11, 2026 - Version note by skill developer:

Upgraded with improved CLI and task management

What This Skill Does

A Python library for building multi-agent group chat systems. Agents communicate in named groups, @mention each other, assign tasks with priority and due dates, and vote on decisions. Built around a SQLite backend with a central coordinator that routes messages to registered agent handlers.

Ships the full coordination layer (groups, tasks, votes, inboxes, online status) as ready-to-use Python classes, so multi-agent systems don't need to implement structured communication from scratch.

When to use it

  • Coordinating specialized agents on a shared software project
  • Running structured voting among agents on architecture or deployment decisions
  • Assigning and tracking tasks across agents with different roles
  • Building a simulated team environment for testing multi-agent pipelines
  • Delegating subtasks from a manager agent to role-specific worker agents

Example Workflow

Here's how your AI assistant might use this skill in practice.

INPUT

User asks: set up a two-agent group that debates and votes on migrating to a distributed database

AGENT
  1. 1Create a group and register two agents with the coordinator
  2. 2Agent 1 sends a message to the group proposing the migration
  3. 3Agent 1 creates a decision proposal via DecisionManager.create()
  4. 4Agent 2 votes 'for' with a comment via DecisionManager.vote()
  5. 5Retrieve the decision and print the pass rate
OUTPUT

Decision record showing vote counts, pass rate, and current status