AI RESEARCH

A Formal Security Framework for MCP-Based AI Agents: Threat Taxonomy, Verification Models, and Defense Mechanisms

arXiv CS.AI

ArXi:2604.05969v1 Announce Type: cross