Last week’s headlines were dramatic. An AI agent pulled off a ransomware attack all by itself. No human at the keyboard. The robots are coming for us.
Calm down.
Sysdig, the cloud security firm that documented the attack, called it “agentic ransomware.” They dubbed it JadePuffer. And yes, the technical execution was impressive. The AI agent broke into a vulnerable server, stole credentials, moved laterally through the network, encrypted files, and wrote its own ransom note. It adapted to obstacles along the way, just like a human hacker would.
But Sysdig’s Michael Clark clarified Monday: a human still set up the operation. Picked the victim. Provisioned the infrastructure — the command-and-control server, the staging server for stolen data. The credentials used to break in? Someone else stole those beforehand and handed them to the operation.
None of this contradicts what Sysdig originally said. The attack itself is still notable. The agent got in through a known bug in Langflow, an open-source tool for building LLM apps. Then it hit a production MySQL server, exploited another known flaw, and gained admin access. It encrypted over 1,300 configuration records. Left a ransom note. Added a Bitcoin address for payment.
Speed was the wild part. The agent fixed a failed login attempt in 31 seconds. It narrated its own reasoning in natural-language code comments the whole way. Sysdig hasn’t disclosed who was targeted.
One lingering question: which AI model was running the show? Sysdig couldn’t identify it. Microsoft researcher Geoff McDonald suspects an open-weight model with safety training stripped out. His reasoning? Frontier labs’ safety layers actually hold up well in his red-teaming tests.
McDonald also warned this means ransomware campaigns are now limited by attacker budget, not human effort. “Thousands or tens of thousands of simultaneous campaigns” could be coming. Though if a human still has to choose each victim and provision infrastructure for every run, that’s still a bottleneck.
Clark says Sysdig hasn’t seen the same operation hit other victims yet. But given how cheap it is to run an AI agent, he expects that to change. The robots aren’t running the show yet. But they’re getting cheaper by the day.
