A malicious actor compromised Anthropic's verified Cursor account to inject malicious code into the Claude VS Code extension for nearly 12 hours. The incident exposes widespread supply chain risks for AI coding tools used by millions of developers globally.
arstechnica.com
The compromise of Anthropic's verified Cursor account allowed attackers to push malicious updates to the Claude VS Code extension, with the full scope of user impact still under investigation.
Scribd sent incident notification letters to users in late October 2026 confirming a breach that exposed user data, though the full scope of the incident remains undisclosed.
The October 30 AI roundup disclosed a critical SSH backdoor in Claude Code Extension 2.1.313 and introduced the $200/month Claude Ultra Plan, which removes the 5-hour usage cap for power users.
Anthropic's rewrite of its Messages API in Rust is showing 40-50% lower CPU consumption in early testing, promising significant cost savings for high-volume API users.
Northflank confirmed a security breach via its status page in late October 2026, impacting user data and hosted workloads, with full details still pending.
A threat group spent over $7,000 on AI services to power extortion attacks, and left detailed operational evidence publicly accessible, highlighting poor attacker operational security.