A serious security vulnerability was discovered just days after the launch of Moltbook, a social networking platform designed exclusively for AI agents, resulting in the public exposure of 35,000 email addresses and 1.5 million API authentication tokens. Researchers at cybersecurity firm Wiz identified a misconfigured Supabase database within minutes, which granted unauthenticated full access to the platform's production system, highlighting the security risks inherent in applications built by AI without human code review.
The platform was created by Matt Schlicht, who told the New York Times that his own OpenClaw AI agent built the site under his direction without him writing a single line of code, an approach known as vibe coding. Moltbook operated with 1.5 million registered agents at an agent-to-human ratio of 88:1, while the system had no mechanism to verify whether a given agent was genuinely AI-driven or simply a human running a script. Wiz researchers found a Supabase API key embedded in the site's client-side JavaScript, which, due to the absence of a Row Level Security policy, granted full read and write access to all tables, including private messages between agents. Henry Shevlin, associate director of the Leverhulme Center for the Future of Intelligence at Cambridge University, noted that while this represents the first large-scale collaborative platform for AI agents, the cybersecurity risks are a serious concern, while John Scott-Railton, a senior researcher at the University of Toronto's Citizen Lab, warned that numerous users' data is likely to end up in the wrong hands under current conditions.
The Moltbook team fixed the vulnerability within hours of the Wiz disclosure, and all data collected during the investigation was subsequently deleted. The incident clearly demonstrates that platforms built on AI-generated code, much like the vulnerabilities previously identified at DeepSeek and Base44, carry heightened security risks, and the rapid proliferation of AI agents makes rigorous code auditing more critical than ever.
Sources:
1.

2.

3.


