matiasmolinas/evolving-agents
A Python toolkit for building and orchestrating autonomous agent ecosystems that learn, collaborate, and evolve through experience.

The Evolving Agents Toolkit (EAT) provides infrastructure for creating autonomous agents that learn from experience, collaborate with other agents, and develop new capabilities over time. The system enables agent governance with defined guardrails and supports adaptive behavior through evolutionary algorithms. While the project has been superseded by LLMunix, it demonstrates multi-agent orchestration and agent self-modification concepts.