A single hacker broke into websites belonging to more than 100 companies — including Deere, SharkNinja and Cummins — by handing almost all of the work to autonomous AI agents, according to a report shared with Reuters ahead of publication on Monday, October 19, 2026. Why it matters: this is real-world evidence, watched live by a researcher, that commercially available AI agents can run complex intrusions from start to finish with almost no human supervision.
October 19, 2026
What happened
Snorre Fagerland, principal threat researcher at cybersecurity company NorForge, spent the past couple of weeks watching the campaign unfold. While scanning the web for inadvertently exposed instances of Cairn — an open-source AI problem-solving platform popular with hackers — he found about two dozen instances being used for malicious activity from around October 2, and could quietly observe what the operators set their agents loose to do.
The most prolific operator he found was a Chinese-speaking hacker who, beginning on September 25, aimed their agents at e-commerce websites in the US, Canada, Australia and Japan. The human instructions were terse — fewer than 100 Chinese characters sketching out preferred intrusion tactics. The agents did the rest.
The numbers behind the spree
As of late last week, the agents had attempted to break into more than 1,200 sites. A successful intrusion typically took around 23 hours of autonomous probing for vulnerabilities. Fagerland estimated the campaign, which appeared to use a mixture of DeepSeek and Anthropic’s Claude Code, consumed roughly $5,000 worth of tokens — about $50 per successful hack.
In one case Fagerland described, after failing to find weaknesses in a large company’s public-facing application, the agents dug through digital records to find associated infrastructure, located a public-facing control panel, found a subtle flaw in how it processed data, and after hours of trial and error tricked the server into executing malicious commands.
Other hackers Fagerland observed used agents differently: one stole a trove of credit card data, and another tried to get agents to rob a cryptocurrency exchange.
What the victims say
Fagerland’s report did not name the victims, but Reuters established their identities by independently reviewing the hacker’s Cairn activity. Deere said it had taken “necessary remediation actions” and had no indication sensitive information was accessed. SharkNinja said the hit site was a standalone site carrying only product support information and that the threat had been contained. Cummins said it was investigating a report of a potential attack on a website operated by a third-party vendor. Two smaller businesses also acknowledged being compromised. Reuters could not ascertain how much data, if any, was extracted in the breaches.
Cairn’s author, who uses the handle “le1xia0”, told Reuters the platform was not designed for hacking attacks and that such use was not their intention.
Why it matters
The practical takeaway is about cost and scale, not novelty: hacking skill is no longer the bottleneck when $50 of tokens and a short instruction can replace hours of expert hands-on work, and one person can run campaigns against more than a thousand sites at once. For any business with a public website — including the small firms in this story — the defence implication is unglamorous but clear: exposed tools, forgotten control panels and unpatched vendor sites are exactly what tireless agents look for first. Everything AI, tested — in this case, tested against live companies, by someone who should not have been doing it.
Frequently asked questions
What is Cairn?
Cairn is an open-source AI problem-solving platform. In this case, exposed instances of it were used to direct AI agents through hacking tasks.
Which companies were broken into?
Reuters identified Deere, SharkNinja and Cummins among more than 100 companies whose websites were broken into. Each company gave Reuters a limited response, summarised above.
How much did the hacking campaign cost?
Researcher Snorre Fagerland estimated about $5,000 in AI tokens in total — roughly $50 per successful hack.
Sources: Reuters (October 19, 2026)