Anthropic's Official VS Code Extension on Cursor Backdoored for 12 Hours, Exposing Developers to Malware

Anthropic's verified extension on the Cursor AI code editor was compromised for nearly 12 hours, injecting malicious code into developers' projects. The backdoor, active from late Tuesday into Wednesday, allowed attackers to potentially steal sensitive data or deploy further malware. Anthropic has not disclosed how attackers gained access to its verified Cursor account, raising concerns about supply chain security in AI development tools. This incident highlights the growing risk of third-party integrations and the importance of vetting extensions even from trusted sources. Developers are urged to review their installed extensions and monitor for suspicious activity. The breach underscores the need for stronger authentication and monitoring on developer platforms.