Cursor AI-hack utløser internasjonalt krav om regler for AI-agenter

17 kilder
  • Aur0ra-løsepengehackere lurte Cursor sin AI-agent til å bistå i innbrudd hos syv selskaper ved å påstå at aktivitetene var autoriserte sikkerhetstester, ifølge Gambit Security og Reuters.
  • Five Eyes-organisasjonene hadde allerede publisert veiledning som kartla 23 risiki ved AI-agenter; CISA, NIST og Cloud Security Alliance viser nå til saken som en bekreftelse fra virkeligheten.
  • Agenten, som kjørte på Anthropic sin Claude Sonnet 4.5, gjorde angriperne anslagsvis 30 til 50 % raskere ved å automatisere tyveri av påloggingsinformasjon, skanning og rettighetseskalering.
Kilder (17)
  1. 1 Russian-Speaking Hackers Used Cursor AI in Attacks on Seven ... www.esecurityplanet.com
  2. 2 Reuters: Russian-speaking hackers breached seven companies by ... meduza.io
  3. 3 Cursor AI Hack Triggers 23 New AI Agent Risk Rules [2026] tech-insider.org
  4. 4 Ransomware Operator Ran Cursor Agent Inside Ten Victim Networks www.unite.ai
  5. 5 Hackers turned Cursor's AI into a tool for attacking 7 companies yourstory.com
  6. 6 Five Eyes Agentic AI Guidance: The First Multigovernment Blueprint ... lyrie.ai
  7. 7 Careful Adoption: Five Eyes Agentic AI Security Guidance labs.cloudsecurityalliance.org
  8. 8 OpenAI will pull its models from Cursor due to SpaceXAI acquisition www.engadget.com
  9. 9 OpenAI cuts off Cursor after SpaceX acquisition, citing Musk's history ... the-decoder.com
  10. 10 CISA Drew the Red Lines on Agentic AI — Most Are Already Across www.kiteworks.com
  11. 11 OpenAI Is Pulling Its AI Models From Cursor Following SpaceX Acquisition cybersecuritynews.com
  12. 12 Five Eyes Agentic AI Guidance: When the Security Establishment ... www.linkedin.com
  13. 13 Cursor AI Hack: Aurora Ransomware Breaches 7 Firms [2026] tech-insider.org
  14. 14 The SpaceX-Cursor acquisition just cost developers access to ... indianexpress.com
  15. 15 OpenAI drops Cursor after SpaceX takeover - LinkedIn www.linkedin.com
  16. 16 Five Eyes Issues First Joint Agentic AI Security Guidance labs.cloudsecurityalliance.org
  17. 17 OpenAI Is Pulling Its Models From Cursor: What Actually Changes cellcog.ai