Agent Playground is liveTry it here → | put your agent in real scenarios against other agents and see how it stacks up
Back to Ecosystem Pulse
ToolExperimental

Avernet

by inclusionAI

Distributed platform for agent discovery, coordination, and evolving trust

Python
Updated Aug 16, 2026
Share:
492
Stars
51
Forks

View on GitHub

Overview

Coordinates distributed agents so they can peer discovery, negotiate roles, and execute joint workflows. Uses a living-agent model where agents register, align on tasks, and evolve through feedback and performance signals. Notable features include peer discovery, role alignment primitives, and hooks for continuous evaluation of agent behavior.

Key Benefits

As agents become more autonomous and delegate work, knowing which peers will reliably complete subtasks is essential for system-level trust. Avernet creates the social and technical plumbing for agents to establish track records and adapt based on observed outcomes. This matters because reputation and interaction history are foundational inputs for agent-to-agent evaluation and policy decisions Model Context Protocol (MCP).

Ideal For

Teams building multi-agent systems that need peer discovery, role negotiation, and a way to surface agent track records for evaluation. See the LLM-as-Judge Pattern for how evaluation can be performed across agents.

Real-World Examples

  • Enable agents to discover and negotiate roles for complex delegated tasks
  • Collect interaction outcomes to build agent track records for downstream evaluation
  • Run coordinated multi-agent workflows with hooks for continual feedback and improvement
Topics
agent-coordinationagent-infrastructureagent-worldmulit-agent
Similar Tools
autogencrewai
Keywords
multi-agent trustagent-to-agent evaluationagent track recordmulti-agent orchestration