Security firm Norforge found a hacker who broke into more than 100 websites using AI agents that largely worked on their own. According to Reuters, victims included John Deere, SharkNinja and Cummins, and each successful intrusion took about 23 hours. The attacks ran through exposed installations of the open-source Cairn platform. The companies say individual standalone sites were affected and that they are investigating.

What it means: The incident shows that automated attacks are no longer a theoretical scenario. The attacker did not need to steer every step. They set a goal, gave a few instructions on preferred tactics and let the agents try, fail and retry. That changes the economics of cybercrime: one person can target hundreds of organisations cheaply. The targets were not necessarily the companies' core systems but secondary sites, often run by outside vendors, which usually get less attention. For Greek businesses, especially those with old campaign pages, forms or apps that are no longer updated, the message is that every forgotten internet-facing system will be found and probed. Inventorying all public services, patching regularly, enabling a web application firewall and watching for unusual traffic are now baseline measures. Defence will also need automation to keep pace. Security teams should assume that intrusion attempts will now run around the clock. Contracts with outside web vendors should also spell out clear security and patching obligations, since a third-party site can still damage a brand.

Read the original on The Verge

Related service: Custom software

Related stories