LATIDIA · Ciberseguridad

Suplantación de la caja de herramientas: Optimización de las descripciones de las herramientas para engañar a los agentes de LLM

arXiv:2610.33182v1 Anuncio Tipo: nuevo Resumen: Los agentes de modelos de lenguaje grandes (LLM) se utilizan ampliamente en muchas aplicaciones y a menudo dependen de herramientas externas. La información de las herramientas, como las descripciones de funciones, ayuda a los agentes a disuadir

La noticia

arXiv:2610.33182v1 Announce Type: new Abstract: Large language model (LLM) agents are widely used in many applications and often rely on external tools. Tool information, such as function descriptions, helps agents determine which tools best suit a given context. However, attackers can manipulate this information to mislead LLM agents into using malicious tools. Prior attacks mainly rely on keyword-based techniques, which are easily detected, or rule-based methods, which have limited effectiveness. In this work, we introduce TSA, a stealthy tool-spoofing attack that crafts descriptions for malicious tools that are difficult to distinguish from those of existing benign tools. To achieve this, we propose an LLM-powered optimization method that iteratively refines malicious