Cursor Says Stolen Fireworks AI Key Led to Trojanized Claude Code Extension Incident
AI-summarised brief · reviewed before publication
Cursor disclosed that a breach of Fireworks AI exposed an access key, which an attacker used to copy data from a read‑only bucket shared with Fireworks. The bucket contained coding session traces from 2025, including unredacted API credentials for less than 0.01 % of customers. Leveraging these credentials, the attacker accessed Cursor’s SaaS tools and, on October 29, 2026, placed a trojanized version of Anthropic’s Claude Code extension into the cache serving Cursor’s marketplace. Up to 5,000 users who downloaded versions 2.1.312 or 2.1.313 received the malicious build, which remained online for about 11 hours before removal. Cursor responded by revoking keys, rotating credentials, disabling services, and notifying affected users, reporting no further malicious activity after November 1, 2026. The incident follows related breaches at Uber and Notion linked to the same Fireworks AI hack.
💡 Why It Matters
- · It shows how a single exposed key can cascade into supply‑chain attacks that compromise developer tools, underscoring the need for stricter credential hygiene and isolation in AI‑driven platforms.
