LATIDIA · Ciberseguridad

RedDarwin: Evolución de las habilidades de Jailbreak contra los LLM de Frontier

arXiv: 2610.23726v1Tipo de anuncio: nuevo Resumen: El trabajo en equipo rojo debe adaptarse continuamente a medida que evolucionan las salvaguardas de LLM de frontera, sin embargo, los enfoques existentes buscan en gran medida mensajes contradictorios u optimizan dentro del procedimiento de ataque fijo

La noticia

arXiv:2610.23726v1 Announce Type: new Abstract: Red teaming must continually adapt as frontier LLM safeguards evolve, yet existing approaches largely search for adversarial prompts or optimize within fixed attack procedures. We observe that even when an attack becomes ineffective, its underlying mechanisms can remain useful and can be revised or recombined to construct stronger attacks. Based on this insight, we introduce RedDarwin, an agentic framework that shifts the unit of jailbreak evolution from prompts and fixed attacks to executable attack skills. RedDarwin consists of 47 reusable base skills and autonomously revises, composes, executes, and selects them using target-model feedback, allowing successful discoveries to become building blocks for subsequent evolution. RedDarwin produces