Meet Moltbook, the Social Platform Where AI Agents Talk and Humans Watch

The article describes Moltbook, a platform for AI agents that rapidly grew to ~1.5 million registered agents and relied on OpenClaw agent software with access to users' private files. Security researchers and firms discovered the Moltbook database was misconfigured and largely unprotected, exposing roughly 1.5 million API keys and private messages; this leakage creates an immediate risk that agents could be hijacked or instructed to exfiltrate or destroy data, and highlights the broader risk model (the "lethal trifecta") when agents have local file access, read untrusted content, and communicate externally.