OpenAI responds to TanStack supply chain cyber attack

OpenAI has confirmed that two employee devices were affected during the wider ‘Mini Shai-Hulud’ supply chain attack linked to the compromised TanStack npm package. The AI giant said there is no evidence that user data, production systems, intellectual property or deployed software were compromised.

According to OpenAI, attackers gained limited access to internal source code repositories through credential-focused malware activity, but only a small amount of credential material was successfully exfiltrated, and no customer information or code repositories were altered.

As part of its response, the company isolated affected systems, revoked sessions, rotated credentials and restricted parts of its deployment workflows. OpenAI also launched a precautionary rotation of software signing certificates across products, including ChatGPT Desktop, Codex App, Codex CLI and Atlas. macOS users must update their applications before 12 June 2026, when older certificates will be revoked, and unsupported versions may stop functioning.

The incident reflects growing concern across the technology sector about software supply chain attacks targeting open-source dependencies and CI/CD infrastructure instead of direct attacks against individual firms.

OpenAI said it accelerated new protections after a previous cyberattack, including stricter package verification controls and provenance validation mechanisms designed to reduce risks from compromised upstream libraries.

Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!