Cisco AI Defense is a security solution for the development and deployment of AI models, applications and agents, including the Model Context Protocol (MCP). It gives visibility of AI assets across cloud environments, scans model files, repositories and MCP servers for risk, validates models through algorithmic red teaming and generates model-specific guardrails. Runtime protection inspects prompts, responses and agent tool calls for threats such as prompt injection and leakage of sensitive data.
Cisco is a US company, so AI Defense falls under US jurisdiction, including the CLOUD Act. According to Cisco, the solution can run in the cloud, in a VPC or on an on-premises AI POD. In VPC and AI POD deployments, application, model and agent data stays in the customer's own environment, and only metadata for management-plane operations is sent to Cisco. The product is not open source. The public documentation names no hosting regions for the management plane, so a European organisation should settle this contractually.