MIRAGE: Stealthy Visual Prompt Injection for Vulnerability Detection in Web Agents

MIRAGE is a novel framework for visual prompt injection targeting web agents under strict spatial constraints, such as an ad slot. It uses diffusion models to generate visually benign adversarial images confined to the attacker's control area. By combining semantic latent manipulation with sparse pixel perturbations, MIRAGE achieves high-efficacy next-action hijacking, demonstrating stealth and realism against existing MLLM agents.

Open the full episode

More from AI Radio