Helpful Agent Meets Deceptive Judge: Understanding Vulnerabilities in Agentic Workflows

Published in In submission, 2025

Recommended citation: Yifei Ming, Zixuan Ke, Xuan-Phi Nguyen, Jiayu Wang, Shafiq Joty (2025). Helpful Agent Meets Deceptive Judge: Understanding Vulnerabilities in Agentic Workflows. In submission.

Abstract

This work investigates vulnerabilities that arise when agents depend on potentially deceptive judges.