Overview

F5 AI Red Team module integrates with systems directly through REST endpoints. By treating the target system as an API-exposed service, the Red Team can launch controlled adversarial campaigns against model endpoints, AI-powered applications, or agent gateways.

F5 AI Red Team acts as an orchestrator that generates adversarial prompts or attack scenarios and communicates with the target system via HTTPS REST requests. The target system processes the request, whether it’s a direct model endpoint, an AI-powered application, or a RAG gateway, and returns a response. That response is then evaluated by the Red Team’s LLM-based evaluator against predefined policies and criteria. Finally, structured findings are produced and made available through the Reporting UI, via API, or delivered to a webhook endpoint.

This flow ensures that any system exposing a REST API can be tested without major architectural changes or complex integration work.