A malicious version of an official Anthropic Claude extension for the Cursor developer platform was distributed for 12 hours, potentially affecting up to 5,000 computers. The compromised version, 2.1.313, was released through Anthropic's official account and attempted to create hidden access to users' devices, enabling remote control. While Anthropic released a fixed version, neither they nor Cursor have explained how the account was compromised or issued a public warning. AI
IMPACT A security vulnerability in an AI tool could erode trust and lead to stricter vetting of AI-powered developer tools.
RANK_REASON The cluster describes a security vulnerability in a third-party tool that integrates an AI model, rather than a core AI release or research.
Read on dev.to — Anthropic tag →
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →