An AI research lab offering an autonomous adversarial agent that probes AI systems in blackbox mode to surface security, logic, and alignment vulnerabilities at scale. Also develops a benchmark framework that measures AI security by quantifying the cost for attackers to exploit a given AI system.