EvoAgentX/Awesome-Self-Evolving-Agents

Curates taxonomy and frameworks for self-evolving AI agents across single-agent optimization (training and test-time behavior), multi-agent systems, and domain-specific applications. Structures agent development techniques—including supervised fine-tuning, reinforcement learning, search-based reasoning, and feedback mechanisms—alongside open-source implementations like EvoAgentX for automated agentic workflow evolution. Links foundational models to lifelong learning paradigms through categorized research papers, code repositories, and conceptual frameworks spanning 2023-2025 advances.

1,939 stars.