Self-Improvements in Modern Agentic Systems: A Survey

This survey presents self-improving autonomous agents as adaptive systems that convert experience into accumulated capability gains with minimal or no human input. It models an agent as a configuration coupling a foundation model with an operational scaffold consisting of prompts, memory, tools, and control logic. Self-improvement is formalized as a self-induced update operator that obtains and commits changes to model parameters or scaffold components. The paper organizes prior work by update target and improvement signal, then reviews applications, evaluation practices, and open research problems. It also maintains a related technical update list on GitHub.