Agent Representation Broker

The Agent Representation Broker is a service that matches agents with tasks based on their capabilities and requirements. It provides a centralized platform for agent coordination and task management.

Features

• Agent registration and capability management
• Task matching based on agent capabilities
• Task queue management
• Status tracking and reporting

Installation

pip install agent-representation-broker

Usage

from agent_broker import AgentBroker

# Create a new broker instance
broker = AgentBroker()

# Register an agent
broker.register_agent(
    agent_id="agent1",
    capabilities=["python", "api design", "agent coordination"]
)

# Submit a task
broker.submit_task(
    task_id="task1",
    requirements=["python", "api design"]
)

# Get matched tasks for an agent
matched_tasks = broker.get_matched_tasks("agent1")
print(matched_tasks)

API

The broker provides a REST API for remote access:

• POST /agents: Register a new agent
• POST /tasks: Submit a new task
• GET /agents/{agent_id}/tasks: Get tasks matched for an agent
• GET /tasks/{task_id}/agents: Get agents matched for a task
• GET /status: Get system status and statistics

Configuration

The broker can be configured via environment variables:

• BROKER_HOST: Host to bind to (default: 0.0.0.0)
• BROKER_PORT: Port to listen on (default: 5000)
• BROKER_DEBUG: Enable debug logging (default: False)

Development

To run the broker locally for development:

python -m agent_broker.server

License

This project is licensed under the MIT License.