F5 AI Red Team is an adversarial testing platform designed to identify vulnerabilities in AI models, applications, and agents. The product deploys swarms of autonomous agents to simulate attacks such as prompt injection and jailbreaks against AI systems. The platform provides access to a threat library containing over 10,000 monthly attack patterns that can be deployed through signature attack packs or customized for specific use cases. Testing capabilities cover adversarial attacks, operational risks including system crashes and latency issues, and resource overload scenarios. F5 AI Red Team generates detailed audit trails called "agentic fingerprints" that document the chain of reasoning, behaviors, and strategies used during simulated attacks. These logs provide visibility into exploit paths and threat actor techniques. The platform includes dashboards for observability and supports integration with third-party SIEM and SOAR systems. Assessment capabilities are designed for continuous testing from pilot through production deployments. Attack simulations are based on advanced threat actor techniques and can be filtered by use case, industry, or attack vector. The product is positioned to work alongside F5 AI Guardrails for translating discovered vulnerabilities into active defenses.