Anthropic denies breach as tampered Claude Code extension hits Cursor users

What happened

Anthropic says nobody breached its own systems. Anthropic says nobody breached its own systems or staff laptops. Anthropic is an artificial intelligence company based in San Francisco, and its products and services include Claude and Claude Code.

What Cursor found The affected versions were 2.1.312 and 2.1.313 for Windows and macOS. Cursor, which SpaceX acquired earlier this year , said Anthropic did not develop the code.

A tampered copy of Anthropic’s Claude Code extension sat in Cursor’s extension marketplace for about 11 hours on Thursday 29 October. The malicious versions tried to plant an SSH backdoor (a hidden way into a system that bypasses its login) on developers’ machines, Cursor said in a security notice.

Data and security The tampered Claude Code extension planted an attacker’s SSH key that uninstalling may not remove.

A security firm that studied the attack says the extension went out through Anthropic’s official account on Cursor.

Neither Anthropic nor Cursor has said how the attacker got access. They were live from about 2:27am to 1:35pm Pacific time.