The rapid advancement of AI technology presents the novel and critical challenge: AI compromise. Cybercriminals are increasingly investigating methods to exploit AI systems for malicious purposes. This includes everything from poisoning development data to bypassing security measures and even deploying AI-powered breaches themselves. The potential