Cursor Says Stolen Fireworks Key Let Attackers Push Backdoored Claude Code Extension to Up to 5,000 Developers
Cursor has confirmed that an access key it issued to inference provider Fireworks AI was exposed in a Fireworks breach and then used by an attacker. The attacker used a credential found in leaked historical coding-agent traces to place trojanized versions 2.1.312 and 2.1.313 of Anthropic's Claude Code extension on the cache serving Cursor's extension marketplace on October 29, 2026. The malicious builds were live for about 11 hours, and as many as 5,000 users who requested them may have received them. The payload attempted to install a persistent SSH backdoor using LaunchAgents on macOS, cron entries on Linux, and scheduled tasks on Windows. Cursor says customer repositories, its production environment, and official app updates were not affected, and users with Privacy Mode enabled were not exposed.
This is an AI-generated summary. ShortSingh links to the original source for the complete article.

Discussion (0)
Log in to join the discussion and vote.
Log in