In brief
- Security firm NorForge says it found more than 20 exposed Cairn control panels likely used in malicious cyber operations.
- Reuters, reviewing one Chinese-speaking actor’s activity, says AI agents broke into more than 100 company websites, including Deere, SharkNinja and Cummins.
- Cairn’s creator says the platform was never a hacking tool. How much data was taken from the named firms is still unclear.
A single operator can now point AI agents at websites and leave them to probe and break in for hours at a stretch, according to researchers who watched the traffic through poorly secured control panels.
That changes the economics of intrusion. NorForge puts an average Cairn task at about $5. For one prolific campaign, its researcher told Reuters the successful breaks worked out at roughly $50 each.
What did NorForge find?
NorForge published Inside Cairn on 19 October. Principal threat researcher Snorre Fagerland scanned the open web for exposed instances of Cairn, an open-source AI platform.
“More than 20 exposed Cairn panels were identified as likely associated with malicious activity,” the report says. Those panels showed how operators pick targets, direct agents, exploit weaknesses and try to keep access.
Cairn, NorForge writes, is “a state-based problem solving engine”: operators set an origin and a goal, and agents work through facts and intents toward that goal. It can be used for many tasks. In the cases NorForge studied, it was being used for penetration testing and break-ins.
The firm also found “a considerable number of apparently legitimate red team operations” left open in the same way, “potentially placing clients at risk.”
Who broke into more than 100 sites?
NorForge did not name the victims. Reuters reported that it independently reviewed one Chinese-speaking hacker’s Cairn activity and identified break-ins at more than 100 companies, including Deere, SharkNinja and Cummins.
That actor was the most active of those NorForge saw. Beginning on 25 September, Reuters reported, their agents went after e-commerce sites in the US, Canada, Australia and Japan. As of late last week they had attempted more than 1,200 sites. Successful intrusions typically took around 23 hours.
NorForge’s own report puts average time to a met goal for its main Chinese-language cluster at a little under 23 hours, across more than 1,100 attack projects on 14 dashboards.
| Claim | Figure | Who says it |
|---|---|---|
| Exposed malicious Cairn panels | More than 20 | NorForge |
| Sites one actor tried | More than 1,200 | Reuters / Fagerland |
| Sites broken into | More than 100 | Reuters |
| Time per successful intrusion | About 23 hours | Both |
| Average cost per task | About $5 | NorForge |
| Cost per successful hack (that campaign) | About $50 | Fagerland via Reuters |
What did the companies say?
Reuters said it could not ascertain how far the breaches led to data being taken. Company responses to Reuters were:
- Deere: took “necessary remediation actions” and had “no indication that sensitive information was accessed.”
- SharkNinja: the hit was a standalone site with product support information only, and the threat had been contained.
- Cummins: “investigating the report of a potential cyber-attack on a website operated by one of our third-party vendors.”
Two smaller businesses also acknowledged being compromised, Reuters reported.
What did the attacks cost?
Keep the two cost figures separate.
NorForge estimates “the average cost of a task to end up around $5 if one includes unsuccessful, early ending projects.” Longer runs could cost more than $100, it says, with considerable uncertainty.
Fagerland told Reuters that the prolific Chinese-speaking campaign appeared to use a mixture of China’s DeepSeek and Anthropic’s Claude Code, and had used “roughly $5,000 worth of tokens, or about $50 per successful hack.”
NorForge’s report names a worker label consistent with Claude Code plus DeepSeek, and another cluster using DeepSeek via OpenRouter.
Is Cairn a hacking tool?
Cairn’s creator says no. The Verge reported that the creator, who uses the handle @le1xia0, said earlier this month that the platform “was never a ‘hacking tool,'” adding that “misuse is a human decision, not a feature of the engine.”
NorForge describes Cairn as a general problem-solving engine that can be pointed at many goals. The open panels it found were being used for attacks; that is use, not the product’s stated purpose.
Our read
This is TSN’s analysis. The story is less about a new exploit class than about scale: one operator, short instructions, and agents that keep going for a day. Exposed panels also mean attackers are leaking their own work. For defenders, the useful signal is the 23-hour window and the cheap retries, not a single branded “AI weapon.”
How we sourced this
- Panel counts, Cairn’s description and the ~$5 average task cost are from NorForge’s report, read in full.
- Victim names and the ~$50-per-success estimate are from Reuters’s exclusive, read via wire reprints after reuters.com and Investing.com blocked direct fetches. The Verge independently matches the victim names and company statements.
- The creator quote is from The Verge. We do not call Cairn a hacking tool.
Sources
- NorForge: Inside Cairn — Open Windows into massive AI-Driven Cyber Operations, 19 October 2026
- Reuters exclusive via Live Mint: How one hacker used AI agents and a few thousand dollars to break into scores of sites, 19 October 2026
- The Verge: A hacker broke into John Deere and SharkNinja with an army of AI agents, 19 October 2026
Related stories
Follow @tsnmediahub on X for breaking AI news as it happens.
Follow on X
