LATIDIA · Ciberseguridad

COPEX: Evaluación comparativa de la robustez de LLM para el contexto adversario en todas las capas del protocolo del contexto del modelo

arXiv: 2610.04378v1Tipo de Anuncio: nuevo Resumen: Los modelos de lenguaje grandes median cada vez más el uso de herramientas en los sistemas de Protocolo de Contexto de Modelo (MCP), donde la influencia adversarial puede entrar a través de instrucciones de usuario, esquemas de herramientas,

La noticia

arXiv:2610.04378v1 Announce Type: new Abstract: Large language models increasingly mediate tool use in Model Context Protocol (MCP) systems, where adversarial influence may enter through user instructions, tool schemas, tool outputs, or protocol messages. Existing benchmarks often evaluate deployed agents, conflating model susceptibility with guardrails, orchestration, and general task capability. We introduce COPEX (COntext Provider EXploitation), a controlled benchmark that isolates the model as an MCP client by fixing the surrounding agent stack and varying only the tool-selecting model. COPEX covers 25 attack types instantiated as 125 scenarios across four entry surfaces: model/agent, client, server/tool, and transport. Across nine models and 3,375 trials, the mean attack success rate is 64.4%, with