Skip to content
NewA calmer, faster ZeroHour

Stolen Fireworks AI key led to Cursor breach and trojanized Claude Code

Original title: Stolen Fireworks AI Key Led to Cursor Data Breach and Trojanized Claude Code Extension

The Hacker News ()3 min read1
Part of a story covered by 19 sourcesFireworks AI breach exposes Notion, Uber and Cursor customer dataMerged summary & timeline
AI summary · x-ai/grok-4.7Summaries by Vani AI

At most 5,000 users may have installed marketplace builds 2.1.312 or 2.1.313, which tried to plant an SSH backdoor.

Cursor said an attacker used an access key issued to Fireworks AI, exposed in that vendor's breach, to copy a read-only bucket of coding-agent traces from sessions not in Privacy Mode. Unredacted credentials from less than 0.01% of customers then opened archived non-production source code, support data, and usage metrics. On October 29, 2026, a stolen credential was used to cache trojanized Claude Code builds for about 11 hours. Cursor rotated credentials, took the marketplace offline, and said it found no further malicious activity after November 1.

Unconfirmed: Uber has not named its provider; TechCrunch reported it is Fireworks AI, and exposure counts are Cursor's figures.

Key facts
  • Most exposed traces dated to 2025 and were eligible for model training.
  • Cursor said Privacy Mode users, customer repositories, and production were not affected.
  • Uber said a provider breach exposed thousands of drivers' licenses; TechCrunch named Fireworks AI.
  • Notion also said its data was stolen in the Fireworks AI incident.
Security assessment · AI, from the article

Impact: Copied agent traces and credentials; trojanized extension tried to install an SSH backdoor.

ProductAffected
Anthropic Claude Code extension on Cursor Marketplace2.1.312 and 2.1.313
What to do
  1. 1.If you installed Claude Code 2.1.312 or 2.1.313 from Cursor Marketplace on October 29, revoke exposed credentials.
  2. 2.Confirm whether agent sessions were outside Privacy Mode and may have been shared with Fireworks.
Analyze this story with Vani AIImpact, who is affected and what defenders should do, in a new chat

From the original article

3 min read at thehackernews.com

The compromise of Anthropic's official Claude Code extension on Cursor's extension marketplace was the result of the theft of an access key that was exposed in the wake of a breach targeting Fireworks AI. "An attacker obtained an access key we had issued to Fireworks," Cursor said in a Friday update. "The key was exposed by Fireworks, and the attacker used it to copy data from a read-only

The Hacker News

thehackernews.com

Read the full article at The Hacker News

A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.