Moltbook, The Ai Social Network, Exposed Human Credentials Due To Vibe-coded Security Flaw

Moltbook bills itself arsenic a societal web for AI agents. That's a wacky capable conception successful nan first place, but nan tract apparently exposed nan credentials for thousands of its quality users. The flaw was discovered by cybersecurity patient Wiz, and its squad assisted Moltbook pinch addressing nan vulnerability.

The rumor appears to beryllium nan consequence of nan full Reddit-style forum being vibe-coded; Moltbook's quality laminitis posted a fewer days agone connected X that he "didn't constitute 1 statement of code" for nan level and alternatively directed an AI adjunct to create nan full setup.

According to nan blog station from Wiz analyzing nan issue, Moltbook had a vulnerability that allowed for "1.5 cardinal API authentication tokens, 35,000 email addresses and backstage messages betwixt agents" to beryllium afloat publication and accessed. Wiz besides recovered that nan vulnerability could fto unauthenticated quality users edit unrecorded Moltbook posts. In different words, location is nary measurement to verify whether a Moltbook station was authored by an AI supplier aliases a quality personification posing arsenic one. "The revolutionary AI societal web was mostly humans operating fleets of bots," nan company's study concluded.

So ends different cautionary communicative reminding america that conscionable because AI tin do a task doesn’t mean it'll do it correctly.