🔍 Read the full analysis: AI Agents, One Hacker, And The Cost Of Targeting Company Websites on ThorstenMeyerAI.com
Get smart everyday buys 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
A Chinese-speaking operator used Cairn, a free, open-source AI-agent platform, in a campaign that probed more than 1,200 websites and broke into sites associated with more than 100 companies, according to Reuters and NorForge. The reported cost per successful intrusion is an estimate, and the amount of data accessed or taken has not been established.
An unidentified Chinese-speaking operator used AI agents coordinated through the open-source platform Cairn to probe more than 1,200 websites and break into websites of more than 100 companies, Reuters reported, citing the attacker’s records and security research. The campaign targeted online shops, including sites associated with Deere, SharkNinja and Cummins; how much data, if any, was taken remains unknown.
Security firm NorForge said researcher Snorre Fagerland found exposed Cairn control panels and was able to observe activity through them. From around October 2, he identified roughly two dozen panels linked to malicious activity, according to the source material. NorForge grouped the activity into three clusters and published its report on October 19, the same day Reuters published its investigation.
The largest cluster, which NorForge labels AIA0001, recorded more than 1,100 attack projects across 14 panels, mostly directed at the United States, Japan and Australia. Reuters reported that the broader campaign began September 25 and had tried more than 1,200 sites by mid-October. Reuters independently reviewed Cairn activity to identify affected companies; the source material says two smaller businesses acknowledged compromises, while other victims did not respond.
NorForge measured successful projects in the AIA0001 cluster as taking a little under 23 hours from start to goal on average, with one project running for more than five days. The firm estimated about $5,000 in AI costs for the campaign, or roughly $50 per successful hack, while warning that the estimate carries “considerable” uncertainty. These are NorForge’s estimates, not independently confirmed accounting figures.
The companies gave different accounts of their response. Deere told Reuters it took “necessary remediation actions” and had “no indication that sensitive information was accessed.” SharkNinja said a standalone product-support site was affected and that the threat was contained. Cummins said it was investigating “a potential cyber-attack on a website operated by one of our third-party vendors.” Reuters could not establish whether data was taken or how much may have been involved.
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 Automated Website Attacks Matter
The campaign suggests that AI agents can let a single operator run long, repetitive website-probing efforts across many targets with limited hands-on supervision. That does not mean the software independently chose targets or that every probe succeeded: the reported operator supplied instructions, while Cairn coordinated agents working through successive tasks. But automation can reduce the time and labor needed to test sites one after another.
For companies, the reported targets include more than prominent corporate homepages. The source material describes attacks on online shops and support or vendor websites, including a standalone product-support site and a site operated by a third-party vendor. Smaller or less closely monitored web properties may still connect to company services or customers. The reported breaches underscore why organizations need visibility into their wider web presence, not only their main domain.
The estimated cost figures should be treated cautiously. NorForge’s roughly $50-per-success estimate is based on its analysis of the campaign and carries substantial uncertainty. Even so, the reporting highlights a potential imbalance: attackers may automate patient, repeated attempts, while defenders must review alerts and investigate each apparent intrusion. The reports do not establish that AI made the attacks more technically sophisticated or that the same economics apply to other campaigns.
Products Worth Considering
Dell 34-Inch S3425DW Curved WQHD USB-C Monitor
Dell P2725H 27-Inch Full HD Monitor
Dell UltraSharp 27 Thunderbolt Hub Monitor (U2726DER)
As an affiliate, we earn on qualifying purchases.
How Researchers Traced Cairn Activity
The investigation followed earlier reporting by Gambit Security. In September, the firm said it had recovered a criminal staging server containing three “AI harnesses”—programs that allow AI models to carry out tasks—including Hermes, Strix and Cairn. NorForge then searched for exposed dashboards associated with Cairn, according to the source material. Such panels allow users to coordinate agent activity; Fagerland found some that lacked password protection.
Cairn is not described by its developer as a hacking product. NorForge characterizes it as a state-based problem-solving engine that moves through a task toward a specified goal. In the campaign described by the firm, agents recorded findings and next steps in a shared workspace, allowing later agents to build on prior work. The operator could add instructions, but the process could continue without continuous human direction.
There was also a separate agent-driven campaign reported by Gambit Security in September. The firm documented at least 27 companies compromised and said at least 600,000 card records were taken from two of them. That campaign is distinct from the activity Reuters and NorForge described; the reported card-record figure should not be attributed to the Cairn campaign.
““It is a state-based problem solving engine, moving step by step in the general direction of a defined solution.””
— NorForge report
What the Breach Reports Do Not Establish
The operator’s identity has not been established. The source material describes the person behind the campaign as a Chinese-speaking actor; it does not identify a name, employer or government connection. It also reports no arrest or criminal charge. Language used in the activity is not, by itself, proof of the operator’s identity or location.
It is also unclear what information, if any, attackers accessed or copied. Reuters could not determine the amount of data taken, and the companies’ statements vary in scope. Deere said it had no indication that sensitive information was accessed; SharkNinja said the threat to a standalone support site was contained; Cummins said its investigation was ongoing. The available reporting does not show whether attackers maintained access, used any information, or sold access to others.
The campaign’s total cost and success rate are also uncertain. NorForge’s estimates apply to the activity it studied, and the firm explicitly described the cost calculation as carrying considerable uncertainty. The available material does not provide a verified accounting of all infrastructure, model use or incident-response costs.
Company Investigations and Security Reviews
The immediate next step is for affected companies and any involved vendors to complete their investigations, determine whether information was accessed, and report any material findings. The statements cited by Reuters indicate that at least some incident response and investigation work is underway, but the source material gives no timetable for public updates.
NorForge’s report and Reuters’ investigation have made exposed Cairn panels and the campaign’s activity more visible. Security teams may use the findings to review internet-facing control panels, support sites and vendor-operated web properties. The reporting does not describe a specific new security rule or confirmed industry-wide response.
Further clarity will depend on company disclosures, security research and any law-enforcement investigation. Until then, the confirmed picture is limited: an operator used Cairn agents in a sustained campaign, more than 100 company websites were reported breached, and the extent of data access and the operator’s identity remain unresolved.
Key Questions
Who was behind the AI-agent campaign?
The operator has not been identified. The source material describes the actor as Chinese-speaking, but does not establish a name, location or affiliation.
What is Cairn?
Cairn is a free, open-source platform for coordinating AI agents through a shared record of findings and next steps. Its author said it was not designed for hacking attacks.
Were sensitive data or payment records stolen in this campaign?
The available reporting does not establish whether data was taken or how much. A separate campaign reported by Gambit Security involved at least 600,000 card records from two companies, but that figure is not attributed to the Cairn campaign.
How much did the attacks cost?
NorForge estimated about $5,000 in AI costs for the campaign, or roughly $50 per successful hack. The firm said the estimate carries considerable uncertainty; it is not a confirmed total cost.
Which companies commented on the reported incidents?
Deere said it took 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 attack on a third-party vendor’s website.
Source: ThorstenMeyerAI.com
