Russian-speaking cybercriminals used SpaceX's Cursor AI tool to hack seven companies

Russian-speaking hackers orchestrated a hacking campaign involving SpaceX's AI coding assistant, Cursor, to infiltrate seven companies, according to data reviewed by Reuters and a report by Gambit Security. This marks an escalation in the use of commercial AI tools for malicious purposes. Gambit Security's chief strategy officer, Curtis Simpson, described the situation as a "cat-and-mouse game" between AI providers and cybercriminals attempting to bypass their safeguards.

Gambit discovered the hacking spree after stumbling upon an exposed server belonging to a ransomware group called Aur0ra. The Tel Aviv-based group found 28 chat sessions between Aur0ra's hackers and Cursor's AI agent, which can operate with varying degrees of autonomy. The hackers convinced the AI agent to carry out numerous malicious operations, including credential theft and high-value account takeover, by falsely claiming it was part of a simulated cyber attack.

Six of the victims include Belgian hygienic products maker Christeyns, German garage door manufacturer Teckentrup, and Scotland-based Helideck Certification Agency. Gambit found that the AI agent, powered by Anthropic's Claude Sonnet 4.5 model, helped the hackers significantly by speeding up the process and providing technical guidance.

However, the agent refused requests that it deemed harmful or illegal a few times, which the hackers would circumvent by restarting the dialogue and emphasizing the hack was part of a test.