bleeping-computer · Crawled Nov 15, 2026

Cursor confirms security breach led to malicious Claude Code extension

2 IoCs
Read original article ↗

AI Summary

Cursor confirmed a security breach stemming from compromised credentials at AI infrastructure provider Fireworks AI, which allowed attackers to distribute malicious versions of the Claude Code extension through Cursor's marketplace. The malicious extension, versions 2.1.312 and 2.1.313, was designed to install persistent SSH backdoors on developers' systems by deploying attacker-controlled SSH keys and executing malicious JavaScript. Up to 5,000 users were potentially exposed during the 11-hour window the malicious extensions were live. The breach also led to the theft of customer data, including unredacted API credentials from AI coding-agent traces, which were used to compromise additional systems.

AI-extracted · verify before operational use

Indicators of Compromise 2 extracted

Type Value Detail
Filename claude-code-2.1.312.vsix Details →
Filename claude-code-2.1.313.vsix Details →