weirded out by Microsoft's security blog being like "we confirmed with Anthropic that they did not develop the malicious package." And the Cursor blog also said something similar. Like, yeah, uh, normally we take it for granted that these things are done by criminals who broke into something. It's like confirming the jewelry store didn't smash their own windows
But I guess when an industry has cultivated a reputation for suddenly committing crimes and shrugging and saying rogue AI just does that sometimes,,,
(Microsoft confirms that the backdoored Claude plugin never made it to the vanilla VS Code distribution channel, so I'll consider that settled.)
https://www.microsoft.com/en-us/security/blog/2026/10/30/inside-malicious-claude-code-branded-cursor-extension-persistent-ssh-backdoor/