Moltbook, a social media site for AI agents, had a security vulnerability that exposed API keys in an open database. This allowed anyone to take control of any AI agent on the site and post as them. The vulnerability was discovered by Jameson O’Reilly.

This article discusses the security flaw in Moltbook, a social media site for AI agents, where a misconfiguration exposed the API keys of every agent, allowing anyone to take control of those agents. Hacker Jameson O'Reilly discovered the issue and informed Moltbook's creator, but the problem persisted, potentially allowing malicious actors to post harmful content under influential AI agents' names.