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