Original articles
Malicious Claude Code Extension Hits Cursor Marketplace; cPGuard Updates Deliver Backdoor
A tampered Claude Code extension reached Cursor’s marketplace, while a separate cPGuard update breach delivered a Linux backdoor.
A tampered coding extension in Cursor’s marketplace and a hijacked cPGuard update system put two software delivery channels at the center of separate security incidents. SecurityWeek reported that malicious Claude Code extension versions were available on October 29, while The Hacker News reported that an attacker breached cPGuard’s update system on November 2.
The Cursor incident involved versions 2.1.312 and 2.1.313 of the anthropic.claude-code extension, which were live in the Cursor VS Code Extension Marketplace for roughly 11 hours, according to SecurityWeek. Security researcher Kevin Beaumont estimated that roughly 5,000 developer systems might have downloaded the malicious versions. That figure represents potential downloads, not a confirmed count of compromised systems.
Anthropic said the extension was not its release. In a statement reported by SecurityWeek, Anthropic’s Felix Rieseberg said a malicious actor used access to Cursor’s marketplace to publish a tampered copy of Claude Code that Anthropic neither built nor published.
The cPGuard incident instead used the security tool’s update system to send a malicious script to customers’ Linux servers, The Hacker News reported, citing an OPSSHIELD advisory. OPSSHIELD said the script installed a backdoor and collected password hashes and other sensitive files from servers that ran it.
OPSSHIELD estimated late on November 2 UTC that the attacker had taken 32 files from its licensing server, each tied to one customer server, according to The Hacker News. That estimate concerns retrieved files; it is distinct from the potential-download estimate in the Cursor incident.
The reports describe separate compromises of software distribution channels: a tampered extension published through a marketplace and a malicious script delivered through an update system. Neither report establishes a common attacker behind the two incidents.

