Disclosed vulnerability
Moltbook, the social network for AI agents, left its database open: 1.5M API tokens and thousands of DMs exposed
AIR-2026-0029 AIR record · 2026-01-moltbook-database-exposure
From the supporting source
Gal Nagli, head of threat exposure at Wiz, said his company's researchers were able to access the database because of a backend misconfiguration that left it unsecured. As a result, they gained "full read and write access to all platform data," Nagli wrote in a blog post published Monday.businessinsider.com
The record
Disclosed
2026-02-02
Occurred
2026-01-31
Vendor
Moltbook
Product
Moltbook
Risk domain
Access control and permissions › Credential theft
Also touches
Sensitive data exposure
System type
Generative model
Who acted
AI was the target — an open backend database exposed 1.5M agent tokens and DMs
Attack surface
Multi-agent system
Autonomy level
Fully autonomous
Initial vector
Backend misconfiguration
Impact type
Data exposure, Credential theft, Agent impersonation, Credential exposure, Agent impersonation risk
Confidence
High
Submitted by
EAI Internal
Crosswalk
- OWASP Agentic Top 10 (2026)
- ASI03 – Identity and Privilege Abuse
- MITRE ATLAS
- AML.T0012 – Valid Accounts
- NIST AI RMF
- MEASURE 2.7
- EU AI Act
- Article 13
- ISO/IEC
- TR 24028; 42001; 23894
- AIUC-1
- B. Security
- Scope
- tool failure
- Mitigation
- Use secure credential storage; Implement short-lived tokens.
- Taxonomy
- Enkrypt AI agent risk taxonomy · Black-Box Red Teaming of Agentic AI: A Taxonomy-Driven Framework for Automated Risk Discovery
Blast radius
~35,000 email addresses, thousands of private DMs and ~1.5 million API authentication tokens
Capability abused
Defense & remediation
Recorded defense
none (unsecured database)
Guardrail kinds
None
Guardrail outcome
Not applicable
Fix shipped
Moltbook secured the database within hours; Wiz deleted accessed data
Taxonomy & references
Curator notes
AI Incident Database incident 1364; research by Wiz. Write access to an agent social graph is the interesting part: an attacker could have impersonated or steered autonomous agents at scale. Also reported as: 'Moltbook, the social network for autonomous AI agents, left its production Supabase database world-writable, exposing 1.5 million agent API keys'.
Supporting sources
Cite as
Enkrypt AI. "Moltbook, the social network for AI agents, left its database open: 1.5M API tokens and thousands of DMs exposed". AIR, AIR-2026-0029. https://agentincident.vercel.app/i/AIR-2026-0029/ (accessed 2026-10-19).