NSA Publishes First Intelligence-Community MCP Security Guidance: 17-Page Threat Model for Agent Automation

NSA·high signal

The NSA's AI Security Center released CSI 'Model Context Protocol: Security Design Considerations for AI-Driven Automation' (U/OO/6030316-26) on May 20 — the first intelligence-community document to address MCP specifically. The 17-page guidance names filtering outgoing proxies, DLP, sandboxing, message integrity, output filtering, and local MCP scans as critical controls, warning of weak authentication, insecure data handling, session hijack, and prompt injection risks. Signals growing government concern about MCP as it becomes enterprise infrastructure.