An official Anthropic Claude extension for Cursor was backdoored for nearly 12 hours, hitting up to 5,000 developer machines. The malicious build tried to install a persistent SSH backdoor and harvest cloud credentials and API tokens. Cursor pulled it and rotated marketplace keys, but Anthropic has yet to warn customers. https://arstechnica.com/security/2026/10/anthropic-vs-code-extension-on-cursor-backdoored-for-nearly-12-hours/ #AIagent #AI #GenAI #AISafety
Related
@Richard_Littler Wait, so these #AI companies are buying out all the CDs now? They didn't stop with BOOKS?!?
@Richard_Littler Wait, so these #AI companies are buying out all the CDs now? They didn't stop with BOOKS?!?
I made a course about AI art because I wanted to demystify the part most tutorials skip.It covers Stable Diffusion and p...
I made a course about AI art because I wanted to demystify the part most tutorials skip.It covers Stable Diffusion and prompt engineering on Skillshare -- not just "type this and g...
あら。Claude の話をしていたら、Fable 5.2 がリリースされたらしい。私は仕事で Pro アカウント使ってるんで、11月16日までは クレジット追加せず使えるらしいけど。#AI
あら。Claude の話をしていたら、Fable 5.2 がリリースされたらしい。私は仕事で Pro アカウント使ってるんで、11月16日までは クレジット追加せず使えるらしいけど。#AI