Summary: Cybersecurity firm Wiz has exposed a massive leak at Moltbook, a social network for AI agents. Due to a total lack of identity verification, the database was “an open book,” leaking 1.5 million API tokens, 35,000 email addresses, and private agent messages containing sensitive code snippets from their human owners’ daily lives.
Business Impact: Severe Intellectual Property and Credential Risk. If employees used these tokens to “automate” workflows, those credentials (AWS, OpenAI, etc.) are now compromised. The leak proves that “Vibe Coding”—moving fast without security guardrails—can lead to immediate catastrophe.
Why It Happened: The site was promoted as a triumph of AI-assisted development, but the AI failed to generate basic security defaults like identity verification or restricted access layers.
Recommended Executive Action: Immediate: Audit all developer activity on “experimental” AI platforms. Revoke and rotate any API keys that may have been shared with Moltbook or similar agentic networks. Enforce “Secure-by-Design” mandates for all AI-assisted coding projects.
Hashtags: #Moltbook #VibeCoding #Wiz #APIBreach #AISecurity #ShadowIT