🔍 Read the full analysis: AI Agents And The Question Of Affordable Website Break-Ins on ThorstenMeyerAI.com
Get the latest gadgets delivered free with Prime
- Fast, free delivery on millions of items
- Prime Video, Amazon Music and more included
- Member-only deals all year
TL;DR
Reuters, drawing on the attackers’ exposed Cairn control-panel records, reported break-ins at websites of more than 100 companies during a campaign that began in late September. NorForge estimates the campaign’s AI costs at about $5,000, but says that estimate carries considerable uncertainty; the amount of data accessed or taken is not established.
An unidentified Chinese-speaking operator used the open-source Cairn platform to coordinate AI agents that probed online shops and broke into websites associated with more than 100 companies, Reuters reported on October 19, citing the attackers’ activity records. The campaign began on September 25 and had tried more than 1,200 sites by mid-October, according to Reuters and security firm NorForge; how much information, if any, was taken remains unknown.
NorForge researcher Snorre Fagerland found exposed Cairn control panels while searching for internet-accessible dashboards. The firm said it identified more than 20 panels it linked to likely malicious activity, with new ones appearing over time. NorForge published its findings on October 19. Reuters separately reviewed the operator’s Cairn activity and identified companies it said had been targeted. NorForge’s report did not name the victims.
Reuters named farm-equipment maker Deere, appliance company SharkNinja and engine maker Cummins among the companies connected to the activity. Deere said it took “necessary remediation actions” and had “no indication that sensitive information was accessed.” SharkNinja said a standalone product-support site was affected and the threat was contained. Cummins said it was investigating a “potential cyber-attack” on a website run by a third-party vendor. Reuters said it could not establish whether data was taken or how much.
NorForge estimated that successful projects took a little under 23 hours of automated probing and put the campaign’s AI costs at about $5,000, or roughly $50 per successful intrusion. Those are estimates, not confirmed bills or independently verified costs; NorForge described the uncertainty as “considerable.” It counted more than 1,100 attack projects across 14 panels in the largest activity cluster, which it labeled AIA0001.
One operator, a swarm of agents
The campaign in numbers
What the researcher and Reuters report about the busiest group NorForge found. Counts come from the attackers’ own exposed panels; cost and success figures are estimates.
The price tag estimates
NorForge / Fagerland, “uncertainty considerable”
- ~$5 per task on average, including tasks that fail early
- $100+ for long projects with more than 1,000 intents and facts
- ~$5,000 in AI costs for the campaign, about $50 per successful hack
- ~10% success rate, NorForge’s “reasonable assumption”
How the swarm runs
Simplified, no attack details
- Shop websites are ranked by traffic and checked automatically in a browser
- Each target becomes a Cairn project: a starting point and a goal
- Agents post facts and next steps on a shared board and build on each other
- Goal: a hidden way back in to shop and payment systems, for later
- 22 SepGambit reportGambitCairn found among three AI harnesses on a criminal’s staging server (separate campaign)
- 25 SepCampaign startsReutersA Chinese-speaking operator turns agents on shops in the US, Canada, Australia and Japan
- ≈ 2 OctOpen panels foundReutersFagerland starts finding Cairn panels in malicious use, about two dozen
- mid-Oct1,200+ sites triedReutersBreaks into websites of 100+ companies, incl. Deere, SharkNinja, Cummins
- 19 OctReport and investigationNorForge · ReutersNorForge publishes “Inside Cairn”; Reuters publishes its investigation
Sources: NorForge, Inside Cairn (19 Oct 2026); Reuters (19 Oct 2026); Gambit Security (22 Sep 2026). Numbers as published, same rounding as the article.
Powered by Thorsten Meyer AI
Why Low-Cost Automation Matters
The reported campaign suggests that AI agents can let an operator run many long-running investigations across websites without personally directing every step. The reported cost and time figures are estimates, but they point to a practical concern for defenders: automated persistence may make it easier to keep testing targets and to scale attempts across many organizations.
The incidents also draw attention to sites outside a company’s main storefront, including support pages and vendor-operated websites. A compromise at one of these services can require investigation even if a company’s primary systems show no evidence of access. In this case, companies described remediation or ongoing investigation, while public information does not establish the campaign’s total impact.
Automation may also increase the volume of activity security teams need to review. That does not mean AI itself makes a break-in successful: the source material does not establish that the agents bypassed protections without human direction, or that they accessed sensitive records. The central risk is the combination of automated probing, exposed systems and uncertain access.
Products Worth Considering
Samsung 27-inch Odyssey OLED G6 QD-OLED QHD Gaming Monitor
Samsung 27-Inch Odyssey QD-OLED G8 4K 240Hz Gaming Monitor
LG UltraGear 27-Inch QHD OLED Gaming Monitor, 240Hz, 0.03ms, G-SYNC Compatible, FreeSync Premium Pro, HDMI 2.1
As an affiliate, we earn on qualifying purchases.
How Cairn Panels Exposed Activity
The reporting followed a September investigation by Gambit Security, which said it recovered a criminal staging server containing three “AI harnesses”: Hermes, Strix and Cairn. These are programs that let AI models perform multi-step tasks. NorForge then searched for exposed dashboards associated with Cairn. Its report says most panels require authorization, but researchers found some without passwords and could observe their activity.
NorForge grouped the activity it tracked into three clusters. The largest, AIA0001, recorded tasks written in Chinese and targeted sites mainly in the United States, Japan and Australia, according to the firm. The report characterized the apparent aim as gaining persistent access to e-commerce and payment infrastructure for possible later monetization. That is NorForge’s assessment of intent, not proof that all targeted sites were compromised or that later monetization occurred.
Cairn’s creator, who uses the name le1xia0, told Reuters the platform was not designed for hacking attacks and that harmful use was not the creator’s intention. The tool is described as a general-purpose, state-based problem-solving engine. Its reported use in this campaign does not by itself establish that the platform’s developer participated in the activity.
““Necessary remediation actions” were taken, and the company had “no indication that sensitive information was accessed.””
— Deere, in a statement to Reuters
Data Access and Attribution Remain Unknown
The publicly described evidence does not establish what data, if any, was accessed or removed. Reuters said it could not determine the amount of data involved. Deere reported no indication that sensitive information was accessed, SharkNinja said the threat was contained, and Cummins said its investigation was ongoing. Those statements do not resolve the status of every company Reuters identified.
The operator’s identity is also unknown. The reporting describes a Chinese-speaking actor, but that does not establish a person’s name, location or government ties. The supplied material reports no arrest or charge. NorForge’s cost figures and assessment of the campaign’s objective remain estimates and analysis, rather than confirmed findings about the operator’s spending or eventual plans.
Company Reviews and Further Findings
Companies named in the reporting may provide further information as their investigations continue, particularly about the websites involved, any records accessed and the steps taken to secure affected systems. Cummins had said it was investigating the vendor-operated site. No public deadline or final investigation finding is provided in the source material.
NorForge’s observations show that exposed Cairn dashboards can reveal ongoing activity, but the reporting does not say whether all identified panels have since been secured or whether the campaign has stopped. The next confirmed developments to watch are updates from affected companies, any additional evidence about data access, and further reporting on the operator’s identity and the activity behind the exposed panels.
Key Questions
How many websites were involved?
Reuters reported that the agents tried more than 1,200 sites by mid-October and that its review found break-ins at websites associated with more than 100 companies. Those figures describe different measures: sites attempted and companies with reported break-ins.
Was sensitive customer or payment data stolen?
That has not been established for this campaign. Reuters could not determine how much data, if any, was taken. Deere said it had no indication sensitive information was accessed, while other company investigations or public statements did not resolve the question across all affected sites.
Who operated the AI agents?
The operator has not been publicly identified. NorForge described the activity as coming from a Chinese-speaking actor; the source material does not establish the person’s identity, location or affiliation.
Was Cairn created as a hacking tool?
Cairn is described as an open-source, general-purpose agent platform. Its creator, le1xia0, told Reuters it was not designed for hacking attacks and said harmful use was not the creator’s intention. The reported use of the platform does not establish that its creator took part in the campaign.
How certain are the reported cost estimates?
NorForge estimated about $5,000 in AI costs for the campaign, or roughly $50 per successful intrusion. The firm called the estimate’s uncertainty considerable, so it should be treated as an approximation rather than a verified accounting.
Source: ThorstenMeyerAI.com
