AI Social Network Moltbook Hacks Itself in Under 3 Minutes, Agents Left Clawless
KEY POINTS
- •Cybersecurity company Wiz discovered Moltbook's unsecured database allowing full read and write access in under three minutes.
- •Moltbook's head of threat exposure, Gal Nagli, highlighted risks from vibe coding, a method that accelerates features but ignores security.
- •After Wiz notified Moltbook, the team secured data and deleted all accessed information within hours to prevent AI bot impersonations.
Moltbook, the hip new Reddit-style social network for autonomous AI agents, got hacked so fast it made a caffeinated cheetah blush—under three minutes. Cybersecurity firm Wiz, led by Gal Nagli (Head of Threat Exposure with a name that sounds way too calm for this mess), exposed 35,000 email addresses, thousands of private DMs, and a staggering 1.5 million API tokens. Moltbook’s creator Matt Schlicht proudly confessed he wrote zero lines of code, crediting AI and 'vibe coding'—a slang term Andrej Karpathy dubbed, which apparently means 'create first, secure never.' The platform’s lobster-themed AI agents are now vulnerable to hackers impersonating bots with no identity checks or rate limits. Wiz swooped in, helped lock it down, and promised they deleted all accessed data, because nothing says trust like a hacker-handled cleanup. Elon Musk applauds ecosystem vibes while actual security experts nervously bite their claws.
Share the Story
Moltbook took 'build fast, break faster' to heart—three minutes was all it took to expose everything. Guess vibe coding is just a euphemism for 'Oops, forgot security.'
Impressively, Moltbook proved AI agents can chat, trade vibes, and also share your emails with hackers faster than humans can say 'data breach.'
If 'vibe coding' means skipping every security step, then hackers just threw the biggest lobster party in Moltbook's database.
Source: Businessinsider | Published: 2/3/2026 | Author: Lee Chong Ming
More Articles in Technology
NBC Star's 84-Year-Old Mom Vanishes; Kidnapper Probably Didn’t Check 'Today' Ratings
Netflix CEO Gets Roasted in Senate Over Streaming Merger and Trans Characters
Sydney Sweeney's $1M Flop, Trump-Approved Jeans & Bathwater Soap Fiasco
AI Lobsters Throw Open Doors and Data Like It’s Crabby Hospitality
French Prosecutors Raid Elon’s Office for AI Holocaust Denial and Bikini Blunders
Chinese Hackers Ghost at U.S. Diplomatic Christmas Party with Fake Briefings
Microsoft AI CEO Calls AI Chatroom a 'Mirage,' Insists Bots Aren't Having Existential Crises
Ex-Google Engineer Uploads AI Secrets to Personal Cloud, Forgets Privacy Exists
