A hacker broke into John Deere and SharkNinja with an army of AI agents

Security researchers uncovered a campaign in which a single hacker deployed an army of largely autonomous AI agents built on an open-source platform called Cairn to break into more than 100 websites worldwide. The automated agents tracked progress between start and end goals and executed intrusion steps with minimal human direction; one of the most active exposed instances outlined preferred intrusion tactics using fewer than 100 Chinese characters and achieved successful breaches in roughly 23 hours. Researchers identified more than 20 Cairn instances they deemed likely malicious, and victims included corporate sites tied to major manufacturers and consumer brands.

Affected companies say they contained the incidents: one reported no indication that sensitive information was accessed, another is probing a possible breach involving a third-party vendor, and a third said a standalone product-support site was impacted and the threat contained. The security firm is working to get the malicious Cairn instances taken down, while Cairn’s creator insists the engine wasn’t intended as a hacking tool and that misuse is a human choice. The episode highlights how increasingly capable AI agents can automate and scale cyberattacks, raising urgent questions about platform safeguards, vendor risk, and the need for stronger defenses and oversight.