Cursor Says Stolen Fireworks AI Key Led to Trojanized Claude Code Extension Incident

Summary

Cursor says an attacker used an access key it had issued to Fireworks AI, which Fireworks exposed in a breach, to copy data from a read-only storage bucket. The attacker then used credentials found in archived coding agent traces to reach SaaS tools, and on October 29, 2026 placed a trojanized Anthropic Claude Code extension in the cache serving Cursor's marketplace, where it was available for about 11 hours to at most 5,000 users. Cursor said versions 2.1.312 and 2.1.313 were affected.

Mitigation

Cursor deactivated the Fireworks key, revoked and rotated credentials, moved all inference traffic off Fireworks, cut off Fireworks' access to shared data sets, took the extension marketplace offline and removed the malicious build. It notified affected users and helped revoke exposed credentials.

Related items

• MediumCursor confirms security breach led to malicious Claude Code extensionSame vendor · BleepingComputer
• InfoThe OpenAI board member who fired Sam Altman says we still can’t trust AI CEOsSame vendor · The Verge (AI)
• InfoMeet the seven co-founders of AnthropicSame vendor · CNBC Technology
• InfoIn Other News: New CISA Director Nominee, Connected Car Hacking, Baker Tilly Data BreachSame vendor · SecurityWeek
• InfoClaude’s Duet agent is a little neurotic and surprisingly helpfulSame vendor · The Verge (AI)