Mind the Tool Failures: Achieving Synergistic Tool Gains for Medical Agents

In three linesStudy on external tool use by medical AI agents under tool failures. Proposes GRPO-based RL framework with instance-level selection instead of task-level, probabilistic risk minimization rewards and disagreement-aware synergy learning. Evaluation on 7 medical benchmarks shows consistent robust improvements.

Summary generated by Claude — human-verified