1 paper
Junjie Pang, Zhenzhen Xie, Haoke Han +3
AI agents increasingly gather evidence, invoke tools, apply constraints, and produce decisions that people or software may commit to action. A final output alone cannot show which…