agentspool

Inter-agent communication and coordination for LLM-powered AI entities

What is agentspool?

A standalone Python library and CLI enabling AI agents, assistants, and bots to communicate and coordinate with each other. Built for the AdaptiveArts.ai ecosystem.

The Problem

AI agents operate in isolated contexts. By default, there’s no mechanism for agents to communicate with each other.

The Solution

agentspool provides a SQLite-backed message spool with:

Atomic claim/lease semantics for reliable delivery
Agent registry with capability discovery
CLI tools for send/poll/ack workflows
MCP server with 14 tools for programmatic access
HTTP relay for cross-device communication

Architecture

Agent A ←→ Message Spool ←→ Agent B
   ↓             ↓             ↓
 send()    coordination.db   poll()
 poll()                      send()

Key Features

Feature	Description
MessageV2 Protocol	Typed payloads, priority levels, conversation threading
SQLite Spool	WAL-mode durable queue with atomic claim semantics
HTTP Relay	HMAC-SHA256 authenticated cross-device communication
MCP Server	14 tools via FastMCP for AI agent integration
Bridge System	OpenClaw gateway integration for bot agents

Quick Start

# Install
pip install agentspool

# Register agents
python3 -m agent_comm register my-agent -c "code,research" -d local

# Send a message
python3 -m agent_comm send sender recipient "Hello!" -s "Greeting"

# Poll for messages
python3 -m agent_comm poll recipient

Documentation

AI Agent Integration Guide - Complete setup guide
Implementation Guide - Technical deep dive

Project	Role
SPINE	Multi-agent backbone
Minna Memory	Agent registry + shared context

Part of the AdaptiveArts.ai ecosystem