Cursor supply chain attack began with Fireworks AI hack
Original title: Cursor Supply Chain Attack Started at Fireworks AI
Tampered Claude Code extension builds 2.1.312 and 2.1.313 were served for about 11 hours to at most 5,000 Cursor users.
Cursor says credentials stolen in hacks of AI inference platform Fireworks AI let attackers take more secrets and publish malicious copies of Anthropic’s Claude Code extension. Builds 2.1.312 and 2.1.313 were available on the Cursor VS Code Extension Marketplace for roughly 11 hours after discovery on October 29; Anthropic was not hacked. A compromised Fireworks-issued Cursor key exposed archived non-production source, support data, metrics, and some credentials users had sent to support. Cursor says access was blocked, the incident is contained, and it has seen no attacker activity since November 1. Notion and Uber also disclosed breaches tied to the Fireworks compromise.
Unconfirmed: Cursor’s account is the source for scope and containment; the exact Fireworks intrusion path is only summarized.
- At most 5,000 users may have received the tampered cached extension.
- Malware was not in the Open VSX marketplace source.
- Uninstalling the extension alone may not be enough, Cursor says.
- Fireworks was hacked twice in October, including a third-party cloud provider.
Impact: Stolen secrets and a malicious Claude Code extension served to users
| Product | Affected |
|---|---|
| Cursor anthropic.claude-code extension | 2.1.312 and 2.1.313 |
- 1.If you installed anthropic.claude-code 2.1.312 or 2.1.313 from Cursor’s marketplace, follow Cursor’s security notice.
- 2.Do not rely on uninstalling the extension alone.
- Vendors
- CursorFireworks AIAnthropic
- Victims
- CursorFireworks AINotionUber
From the original article
2 min read at securityweek.com
The attacker stole an access key from Fireworks and used it to exfiltrate credentials that enabled them to push malicious Claude Code extension copies. The post Cursor Supply Chain Attack Started at Fireworks AI appeared first on SecurityWeek.
By Ionut Arghire · securityweek.com
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.