LATIDIA · Ciberseguridad

Autorización en tiempo de ejecución de subobjetivos autogenerados en agentes de IA que utilizan herramientas de Long-Horizon

arXiv: 2610.04975v1Tipo de anuncio: Cross Resumen: Los agentes de IA que utilizan herramientas de horizonte largo crean subobjetivos, replanifican, delegan trabajo y redactan resultados hermanos. Las comprobaciones de permisos por herramienta no pueden establecer que un objetivo cambiante gr

La noticia

arXiv:2610.04975v1 Announce Type: cross Abstract: Long-horizon tool-using AI agents create subgoals, replan, delegate work, and compose sibling results. Per-tool permission checks cannot establish that a changing goal graph remains within the principal-approved task. We address this authorization gap in a finite structured domain with one principal and one authorization root. Each proposed goal-graph mutation carries a version-bound witness that its continuation traces, resources, obligations, invariants, and closing condition refine the active root contract; every protected effect is rechecked at an atomic commit boundary. Free-form goal text supplies no authority. We prove trace-policy and modeled forbidden-state preservation under explicit mediation, abstraction, freshness, and atomicity assumptions, plus conditional root-success preservation, a separation