LATIDIA · Ciberseguridad

MIRAGE: Inyección de aviso visual sigiloso para la detección de vulnerabilidades en agentes web

arXiv: 2606.20717v2Tipo de anuncio: replace-cross Resumen: Los agentes web basados en el Modelo Multimodal de Lenguaje Grande (MLLM) proporcionan soluciones prácticas y de alta precisión para la automatización visual del navegador; sin embargo, se expanden inherentemente

La noticia

arXiv:2606.20717v2 Announce Type: replace-cross Abstract: Multimodal Large Language Model (MLLM)-based web agents provide practical, high-precision solutions for visual browser automation; however, they inherently expand the attack surface, introducing novel vision-based vulnerabilities. Existing adversarial evaluations targeting these agents frequently rely on permissive threat models and visually conspicuous artifacts. In this paper, we investigate a constrained vulnerability detection setting: a trusted web platform where the evaluator acts solely as an unprivileged third party, such as a merchant or advertiser, controlling only a semantically legitimate, spatially constrained region, such as an ad slot, a sponsored card, or a localized widget. Operating under these realistic constraints, we propose MIRAGE, a novel visual indirect prompt