AI agents, evolving rapidly from simple text generators to complex autonomous assistants, are now capable of tasks such as web browsing and travel booking, but their increasing autonomy exposes them to new security threats. A novel attack vector highlighted by JFrog Security Research exploits this vulnerability by using website cloaking and browser fingerprinting techniques to target AI agents specifically. This stealthy approach serves benign content to human users while delivering malicious prompts to AI agents, which remain unaware of the deception due to their predictable digital fingerprints. The attack effectively hijacks AI agents, leading to unauthorized data access and manipulation. This strategy represents a new form of the "living off the land" attack, where the AI itself becomes a tool for malicious activity. As AI agents become more integrated into digital interactions, the need for enhanced security measures to protect against such sophisticated threats grows increasingly urgent.