Introducing Agent Harness Testing in Cisco AI Defense
Cisco, Tuesday, June 16th, 2026
Cisco AI Defense adds Agent Validation, an autonomous red-teaming capability for agentic AI systems.
Cisco launched Agent Validation, a new evaluation capability within AI Defense: Explorer Edition built specifically for agentic AI systems. The tool performs autonomous red teaming to test agent-specific attack surfaces, including tool routes, indirect content channels, and persistent state modifications.
These are areas traditional chat-based evaluations typically miss. It aims to help organizations secure AI agents before deployment.