🔍 Read the full analysis: Could A Swarm Of AI Agents Change How One Hacker Targets Websites? on ThorstenMeyerAI.com
Get business pricing on monitors, keyboards and dev gear
- Business-only prices and quantity discounts
- Tax-exempt purchasing
- Multiple users, one account, clear invoices
TL;DR
A Chinese-speaking operator used the open-source Cairn platform to coordinate AI agents that probed more than 1,200 websites and broke into sites associated with more than 100 companies, according to Reuters and security firm NorForge. The reported count comes from the attackers’ own records; the amount of data accessed or taken has not been established.
A Chinese-speaking operator used the open-source platform Cairn to coordinate AI agents that probed online shops from September 25 and broke into websites associated with more than 100 companies, according to Reuters’ review of attacker records and research by NorForge. The targets were in the United States, Canada, Australia and Japan; how much information, if any, was accessed or taken remains unknown.
By mid-October, the agents had tried more than 1,200 sites, according to the campaign records described by Reuters and NorForge. Reuters identified Deere, SharkNinja and Cummins among companies whose websites were affected. NorForge’s report, published October 19, grouped related activity into clusters and counted more than 1,100 attack projects across 14 panels associated with its largest cluster, which it named AIA0001. That figure describes projects recorded on those panels, not a count of successful intrusions.
NorForge researcher Snorre Fagerland found exposed Cairn control panels while searching internet-indexed servers. The firm said it identified more than 20 panels it linked to likely malicious activity and was able to observe some operations close to real time. Reuters separately reviewed Cairn activity to identify affected companies. Two smaller businesses acknowledged being compromised, while other contacted victims did not respond, Reuters reported.
NorForge measured successful projects in the largest cluster as taking a little under 23 hours from start to goal. Fagerland estimated the campaign’s AI costs at roughly $5,000 overall, or about $50 per successful intrusion, but NorForge cautioned that the estimate carried considerable uncertainty. These are the researcher’s estimates, not independently verified accounting figures.
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
How Automation Changes the Workload
The campaign illustrates how an operator can delegate repeated probing to multiple AI agents rather than work through targets one by one. NorForge described Cairn as a general problem-solving engine, not a hacking product; in this reported campaign, agents shared findings and next steps through a common control panel. That structure can let a small number of people maintain activity across many sites over extended periods.
For website operators, the reported targets and company statements point to the importance of checking support sites, vendor-run pages and older web infrastructure as well as primary storefronts. Reuters could not establish whether sensitive data was taken, so the incident count should not be treated as a measure of financial loss or customer exposure. Still, a campaign that generates many automated probes can add pressure to security teams, which must distinguish routine scanning from attempts that require investigation.
NorForge’s cost estimate also suggests why defenders may face sustained activity: the researcher put the AI expense at around $50 per successful intrusion, while stressing the estimate was uncertain. That figure does not include all possible costs or establish the attacker’s profit. Its relevance is narrower: according to the report, automation made it possible to spend substantial time on each successful project without constant human attention.
Popular right now
Ninja Foodi DualZone 10-Qt Air Fryer, 6-in-1, Black (DZ302)
KitchenAid Artisan 5-Qt Tilt-Head Stand Mixer – Almond Cream (KSM150PS)
Vitamix 5200 Professional-Grade Blender, 64 oz. Container, Black
As an affiliate, we earn on qualifying purchases.
From Exposed Panels to Targeted Shops
The investigation began after Gambit Security’s September report on a separate campaign. Gambit said it recovered a criminal staging server containing three AI harnesses, including Cairn, software that coordinates model-driven tasks. NorForge then searched for exposed dashboards associated with Cairn. Its report says most panels required authorization, but some did not; Fagerland found panels that could be viewed without a password.
NorForge identified roughly two dozen panels it associated with malicious activity from around October 2, according to Reuters’ account. New panels continued to appear, and the firm organized the activity into three clusters before publishing its findings on October 19. NorForge’s report named no victims; Reuters said it identified companies by independently reviewing Cairn activity and contacting them.
The platform’s developer, who uses the name le1xia0, told Reuters that Cairn was not designed for hacking and that harmful use was not the developer’s intention. Cairn’s broader design matters when interpreting the case: the reported abuse involved a general-purpose agent platform, not a tool described by its author as built specifically for cyberattacks.
“It is a state-based problem solving engine, moving step by step in the general direction of a defined solution.”
— NorForge report
Data Access and Attribution Gaps
Reuters said it could not establish how much data, if any, was taken. Deere told Reuters it had no indication that sensitive information was accessed. SharkNinja said a standalone product-support site was affected and that the threat had been contained. Cummins said it was investigating a potential attack on a website operated by a third-party vendor. These statements describe different company responses and do not establish the full scope of exposure across all reported targets.
The operator’s identity is not known. NorForge described the actor as Chinese-speaking, which does not establish nationality or location. The campaign’s recorded activity, the number of successful intrusions and the AI-cost estimate also rely on attacker records or researcher analysis; the source material does not provide an independent audit of those records. No arrest or criminal charge was reported.
A separate Gambit Security report described another agent-driven campaign involving at least 27 companies and at least 600,000 card records taken from two of them. That is a distinct campaign, not evidence that payment-card data was taken in the Cairn activity described here.
Company Reviews and Security Follow-Up
Companies affected by the reported activity will need to determine which systems were reached, whether any data was accessed and whether the intrusions left persistent access behind. The company statements cited by Reuters indicate that remediation or investigation was under way at Deere, SharkNinja and Cummins, but the source material gives no timetable for final findings.
Further reporting could clarify how many of the more than 100 company websites suffered confirmed unauthorized access, what information may have been exposed and whether investigators can identify the operator. NorForge’s report and Reuters’ investigation provide a view of the campaign through exposed control panels and attacker records; neither account, as described in the source material, resolves those outstanding questions.
Key Questions
What did the AI agents do?
According to Reuters and NorForge, agents coordinated through Cairn probed online shops and were involved in successful intrusions at websites associated with more than 100 companies. The reports do not establish that every site the agents tried was breached.
Who was behind the campaign?
The operator has not been publicly identified. NorForge described the actor as Chinese-speaking; that description does not establish the person’s nationality or location.
Was customer or payment information stolen?
It is not known whether data was taken in this campaign. Reuters could not determine the amount, if any, and Deere said it had no indication that sensitive information was accessed. Gambit Security’s separate report about stolen card records concerned a different campaign.
Was Cairn built for cyberattacks?
No, according to its developer, le1xia0, who told Reuters that Cairn was not designed for hacking attacks. NorForge described it as a general-purpose, state-based problem-solving engine that can coordinate AI agents.
How reliable is the estimate of the campaign’s cost?
NorForge researcher Snorre Fagerland estimated roughly $5,000 in AI costs for the campaign, or about $50 per successful intrusion. NorForge described the estimate as carrying considerable uncertainty, and it is not an audited figure.
Source: ThorstenMeyerAI.com
