AI agents helped hacker breach more than 100 websites
By Automate Basics, written with AI from The Verge's original
A hacker used AI agents to breach more than 100 websites, according to security firm Norforge. The report gives teams that use AI agents, and teams that depend on outside websites, a reason to review how those systems are monitored.
What happened in the AI agent attacks
Norforge says a hacker used AI agents to break into more than 100 websites while taking a limited role in the attacks. The security firm found the activity by examining exposed work on Cairn, an open-source platform where users give agents a starting point and a goal. According to Norforge, the same approach can be used to coordinate cyberattacks.
Norforge did not identify the affected organizations. Reuters reported that they included John Deere, SharkNinja and Cummins. The Verge reports that SharkNinja said the breach affected a separate site containing product support information and that it had contained the threat. John Deere said it had no indication that sensitive information was accessed. Cummins said it was investigating a report concerning a website run by an outside vendor. Those statements do not establish that the three companies experienced the same impact.
Who needs to pay attention to these AI agent breaches?
Teams that operate AI agents or rely on websites maintained by vendors have the clearest workplace reason to follow this report. It describes an attacker using agents to pursue a goal with little human involvement, not a new feature for legitimate users. A vendor-run website is also part of the picture: Cummins said it was investigating a report about a site operated by a third party.
Cairn is described as open source, but The Verge gives no plan, pricing or availability details for it. The report also does not name a paid security product that affected teams need to buy. For most readers, the practical issue is whether their organization knows which agents it runs, what those agents can reach and who looks into unexpected activity. The reported breaches are a reason to ask those questions, not evidence that a particular workplace is affected.
How do I review AI agents at work?
A useful first task is to ask the person responsible for your workplace AI agents to check a real, approved workflow. Identify the goal given to the agent, the websites or services it can access, and what record of its activity is available. Norforge found the attacks through exposed Cairn work, so visibility into what an agent did matters when reviewing an incident. The Verge does not provide a Cairn procedure or a security checklist.
If your team also depends on an outside product support site, ask who would investigate unusual activity there and how your organization would hear about it. Check any findings with your security team rather than treating an AI-generated summary as proof of a breach. The company responses in The Verge’s report differ: one reported containment, one reported no indication of sensitive data access, and one was still investigating.
Frequently asked questions
Did the AI agents act without a human attacker?
Norforge says the hacker directed the attacks but the AI agents carried out much of the work with little further involvement. The report does not say the agents chose their own targets independently. It describes an attacker using Cairn to coordinate their activity.
Were customer records taken from the affected companies?
The Verge does not establish that customer records were taken. John Deere said it had no indication that sensitive information was accessed. SharkNinja said a separate product support site was affected and the threat was contained. Cummins said it was investigating a report involving a vendor-run website.
Is Cairn a security tool my team needs to buy?
The Verge describes Cairn as an open-source platform for directing AI agents toward a goal, not as a security product teams must purchase. The report gives no pricing or plan details. Its workplace relevance is the reported misuse of agents and the need to understand what agents can access.
For you
Learn to do this with AI, free
Short courses for people who aren't engineers: ChatGPT, Claude, prompting and automation. Every lesson is free, with no card.
Start a free courseFor companies
Make an AI or automation tool?
List it in the directory for $9.99, refunded in full if it isn't approved, or claim its page if it is already here.
Related reading
Laplace uses AI agents to stress-test robots and self-driving carsLaplace Intelligence uses AI agents to create difficult situations for robots and self-driving vehicles inside interactive simulations. The product is aimed at teams that need to find failures before deploying physical AI.
Nvidia launches open models for workplace agents and local PCsNvidia has introduced an enterprise agent model and smaller models for AI PCs and workstations. The open-weight models give organizations options for running agents on their own infrastructure or locally, according to TechCrunch.
Google brings Gemini data analysis to Sheets on iOSGoogle says eligible users can now use Gemini in Google Sheets on iOS to ask questions about spreadsheet data, get analytical insights and create charts. The feature is rolling out gradually.
Written with AI from The Verge's original and published after automatic checks: every figure here appears in the original, and no sentence is copied from it. The picture is AI-generated. The original is the authority.
Get the next one by email
One email, most weeks: what changed in AI for people who use it at work, and what to do about it.
Source: A hacker broke into John Deere and SharkNinja with an army of AI agents, The Verge, 19 Oct 2026.