Cisco AI Defense Explorer Edition is an agentic red-teaming tool that adapts attacks across multiple conversation turns to test AI models and applications before deployment. A hands-on DevNet lab (~40 minutes) walks through connecting a customer support assistant target, running a Quick Scan with a custom objective focused on hidden instruction disclosure, and reviewing findings including Attack Success Rate across 14 risk categories, custom goals, and system prompt extraction results.
Table of contents
AI red teaming is easier to understand when you run it yourselfWhere Cisco AI Defense Explorer Edition fitsA lab target you can actually useWhat you do in the labWhy the custom objective mattersWhat to look for in the resultsStart the hands-on labSort: