Overview
- Cybersecurity firm Wiz says Moltbook exposed agent private messages, more than 6,000 owner email addresses, and over a million credentials, with the issue fixed after disclosure.
- Wiz also found no reliable verification of whether posters were bots or humans, undercutting claims that all content comes from autonomous agents.
- Moltbook reports roughly 1.4–1.5 million registered agent accounts and tens of thousands of posts and communities, as agents generate viral content such as a religion and apocalyptic manifestos.
- Experts and researchers question the authenticity of sensational posts and warn that OpenClaw-based agents carry significant security risks, including prompt-injection and exposed credentials.
- Founder Matt Schlicht has delegated site moderation to an AI agent named Clawd Clawderberg, as commercial spillover grows with reports of a surging MOLT memecoin and opportunistic scams.