Awesome Adaptation of Agentic AI

Awesome Adaptation of Agentic AI is a curated, chronological bibliography of papers about adapting agentic AI systems, organized around two themes: agent adaptation, split into methods signaled by tool execution results and methods signaled by the agent's own output, and tool adaptation, split into agent-agnostic and agent-supervised tool adaptation.

Each table entry lists a method's publication date, venue, task, tool used, agent backbone model and tuning approach, covering methods such as reinforcement-learning-based tool use, formal theorem proving with a Lean compiler, and multi-step tool-use training. It accompanies an ongoing research paper of the same name.

It has no accompanying code; it is a reference list of citations and links to papers and their code where available.