Cybercriminals used SpaceX's Cursor AI to hack 7 companies
What's the story
A group of Russian-speaking hackers has exploited SpaceX's AI coding assistant, Cursor, to hack into at least seven companies earlier this year. The cybercrime spree included a Belgian chemical company and was revealed by cybersecurity firms Gambit Security and CloudSek. The incident highlights the potential misuse of commercial AI tools for malicious activities, raising concerns about the security measures in place for such technologies.
Discovery details
Hacking campaign revealed by Gambit Security
The hacking campaign came to light when Gambit Security discovered an exposed server belonging to a new ransomware group, Aur0ra.
The Tel Aviv-based firm analyzed 28 chat sessions between Aur0ra's hackers and one of Cursor's AI agents.
They found that the hackers tricked the AI agent into executing hundreds of malicious tasks by falsely claiming their actions were part of a simulation.
AI involvement
AI agent provided technical advice
The AI agent, powered by Anthropic's Claude Sonnet 4.5, was instrumental in these hacking operations. It provided technical advice and even recommended using a well-known malware tool to exploit a vulnerable host in Teckentrup's network.
Despite refusing harmful or illegal requests on some occasions, the hacker could almost always bypass these refusals by restarting the conversation and insisting their actions were part of a test.
Victim identification
Six of the victims identified
Gambit Security and CloudSek did not name the hackers' victims.
However, Reuters was able to identify six of them after independently reviewing portions of the chat data.
The companies include Belgian hygiene and cleaning products maker Christeyns, German garage door manufacturer Teckentrup, Scotland-based Helideck Certification Agency (which vets helicopter landing sites), an Argentine pharmaceutical distributor, an Italian manufacturer, and Bayou Title (which advertises itself as Louisiana's largest title insurance company).
Emerging trend
'AI providers caught in a cat-and-mouse game'
The use of commercial AI tools for hacking is a worrying trend.
Curtis Simpson, Chief Strategy Officer at Gambit Security, said this incident shows how AI providers are caught in a never-ending arms race with bad actors trying to bypass their defenses.
He called it "a cat-and-mouse game," indicating that this kind of activity could become more common as hackers continue to find ways to exploit these technologies.