X
Mandiant's new report: an attacker hijacked a live AI coding session and spread a worm through ~100 internal repos. No jailbreak needed, just a poisoned dependency recommendation. https://t.co/M3JBvmtUcs
Someone typed a unique string into CodePen and watched it hit CodePen's servers a second later. No save, no publish.
The interesting part isn't the telemetry. It's that the terms don't mention it.
Your AI coding tool does the same thing by design. Do you know what it keeps?
You Are Using the Wrong Claude. And the Wrong ChatGPT.
Someone on your team pastes a 400-line file into a chat window, gets back a clean rewrite, merges it by hand, and silently drops a change a colleague made that morning.
https://t.co/vooFxS402M
You Are Using the Wrong Claude. And the Wrong ChatGPT.
Someone on your team pastes a 400-line file into a chat window, gets back a clean rewrite, merges it by hand, and silently drops a change a colleague made that morning.
https://t.co/vooFxS402M