A report from Google reveals that cybercriminals utilized a multi-agent AI framework to successfully compromise thousands of credentials within six hours. This threat was traced back to a financially motivated individual who initially breached an organization's cloud infrastructure before deploying the AI agents.