LATIDIA · Ciberseguridad
AutoPI: «red teaming» adaptativo guiado por taxonomía para la inyección de «prompts» en agentes LLM
arXiv:2611.24313v1 Tipo de anuncio: nuevo Resumen: La inyección de «prompts» es el principal riesgo de seguridad no resuelto para los agentes de LLM, y la comunidad ahora coincide en que la solidez debe medirse frente a atacantes adaptativos en lugar de
WhatsApp ↗Telegram ↗
La noticia
arXiv:2611.24313v1 Announce Type: new Abstract: Prompt injection is the dominant unsolved security risk for LLM agents, and the community now agrees that robustness must be measured against adaptive attackers rather than fixed benchmark injections. Many adaptive red teamers exist, and most are built to maximize attack success rate (ASR). Yet ASR counts how often a target breaks, not in how many different ways. A red teamer can achieve a high ASR by reusing a few techniques that work, leaving the others untested and absent from the attacks used to harden models. We present AutoPI, an adaptive red teamer designed to find many different ways to break a target, not just