Chinese hacker attacks more than 100 company websites using artificial intelligence agents: Deere and SharkNinja are among the major brands affected by the attack
follow us
Add us as a preferred source on Google
- A hacker used artificial intelligence agents to attack more than 1,200 websites and compromised more than 100 in less than a month.
- Autonomous agents tested vulnerabilities, adapted tactics, and established access with minimal human intervention.
- AI-powered attacks are making large-scale website compromise faster and cheaper, researchers warn
In less than a month, a single Chinese hacker managed to break into more than 100 company websites, doing almost none of the work himself. Instead, he commissioned AI agents to do it for him and paid them around $5,000 in tokens for the job.
The campaign was detected by cybersecurity researcher Snorre Fagerland of NorForge, who was searching for exposed Cairn installations when he discovered that approximately two dozen were being used for malicious purposes. This allowed him to see first-hand an active cybercriminal operation.
Cairn is described as an open source, general-purpose AI problem-solving engine designed to help autonomous agents find solutions to complex tasks. Instead of focusing on a single application, Cairn provides a framework that defines goals, explores different approaches, and then takes actions to try to solve the problem. According to numerous media outlets, Cairn is quite popular among the cybercriminal community (although it is not malicious by design).
Latest videos ofTechnologyRadar
When security becomes a problem that needs a solution
In this case, the “problem” Cairn was trying to solve for the attacker was how to run code on a website without logging in. They were testing SQL injection, file uploading, and insecure handling of stored objects. The goal, according to the researcher, was to establish persistent access to purchasing and payment systems that could then be used to steal money.
Fagerland said the attacker used a combination of models, probably Claude Code with a DeepSeek model, although it is very difficult to determine. The instructions themselves were very short, usually less than 100 Chinese characters, and in less than a month, agents attempted to break into more than 1,200 sites.
On average, agents snoop around a website for a day (23 hours) before logging in or moving on. With about 100 successful thefts, Fagerland says the hacker paid $50 per token on average, per breach. However, this is an estimate and not a confirmed billing.
The analyst did not name the victims. In its report, Cybersecurity News said the victims were mainly online stores in the United States, Canada, Australia and Japan, and Reuters managed to confirm some targets by “independently reviewing hacker activity on Cairn”: among them Deere (agricultural equipment manufacturer), SharkNinja (appliance company) and Cummins (engine manufacturer). He added that two “smaller companies” also acknowledged being compromised but did not respond to the agency’s questions.
Impressive technique
Deere said Reuters took “necessary corrective action” and added that “there was no indication that confidential information had been accessed.” SharkNinja said the breach affected a standalone site that only contained product support information and that the threat had been remediated. Cummins just said it was investigating the reports.
So far, no one knows if the attacker managed to steal sensitive data from his victims and, if so, how much. The investigator reportedly found “hundreds” of secrets exposed in the open panels and was working with Reuters to let the victims know.
Fageraland told Reuters he was “shocked” by some of the techniques officers used to compromise their victims. Apparently, after reaching a dead end against one of the targets, the AI agents scanned digital records for associated infrastructure. When they found a public dashboard, they discovered a “subtle weakness” in the way it processed data. It took “hours of trial and error” before finally being able to execute a malicious command on the website’s server.
“It’s not easy to find solutions like that,” Fagerland said. “These are not scripts for children.”
Apparently, AI attackers are the new normal. Business owners are advised to patch public applications, restrict access to administration panels, and keep an eye out for unexpected files in web directories.
Through Tom Hardware
The best antivirus for all budgets
Follow TechRadar on Google News and add us as a preferred source to receive news, reviews and opinions from our experts in your feeds.
Post Comment