LATIDIA · Ciberseguridad

EvoSafeHarness: Evolución de arneses específicos de modelo y dominio para agentes de seguridad

arXiv: 2609.05903v2Tipo de anuncio: reemplazar Resumen: Los agentes del modelo de lenguaje grande (LLM) pueden traducir las solicitudes de lenguaje natural en acciones del mundo real, creando riesgos de seguridad tanto de inyecciones rápidas indirectas como directas

La noticia

arXiv:2609.05903v2 Announce Type: replace Abstract: Large language model (LLM) agents can translate natural-language requests into real-world actions, creating safety risks from both indirect prompt injections and directly harmful requests. Existing system-level safety harnesses are typically designed by experts and applied across heterogeneous models and domains, despite substantial differences in model behavior, utility tolerance, and domain-specific safety requirements. We present EvoSafeHarness, a framework for automatically synthesizing deployment-specific safety harnesses for frozen LLM agents. Given a target model and domain specification, EvoSafeHarness jointly evolves a natural-language policy and executable enforcement logic using behavioral feedback from the target model. Each candidate is additionally evaluated through fresh-context adversarial review to filter rules that depend