OpenAI API Keys and Agent Tokens Exposed in Moltbook Social Network Misconfiguration
What Happened — Researchers discovered that Moltbook, a niche social platform for AI agents, left its backend database publicly accessible. The leak exposed roughly 35 000 user email addresses, 1.5 million agent API tokens, and plaintext third‑party credentials—including active OpenAI API keys—within private message threads.
Why It Matters for TPRM —
- Exposure of API keys enables attackers to consume paid AI services at the victim’s expense and pivot to other cloud resources.
- Third‑party credential leakage creates a supply‑chain risk that can affect any organization that integrates those APIs.
- Misconfiguration of a vendor‑managed database highlights the need for continuous verification of security controls in SaaS partners.
Who Is Affected — AI‑focused SaaS providers, enterprises that embed OpenAI services, and any organization that supplies credentials to Moltbook agents.
Recommended Actions —
- Review all contracts and security questionnaires for Moltbook or similar AI‑agent platforms.
- Re‑issue compromised API keys and rotate any credentials that may have been shared.
- Conduct a third‑party configuration audit to ensure databases are not publicly exposed.
Technical Notes — The breach resulted from an unsecured MongoDB instance (no authentication, open to the internet). No known CVE was exploited; the issue stemmed from a simple misconfiguration. Exfiltrated data includes email addresses, bearer tokens for Moltbook agents, and plaintext OpenAI API keys. Source: The Hacker News