Toxic Combinations: When Cross-App Permissions Stack into Risk

Summary: On January 31, 2026, researchers disclosed that Moltbook, a social network built for AI agents, had left its database open, exposing 35,000 email addresses and 1.5 million agent API tokens across 770,000 active agents. Some private messages held plaintext third-party credentials, including OpenAI API keys shared between agents, stored in the same unencrypted table as the tokens needed to hijack the agent itself.
What's compelling: The exposure is a toxic combination: a permission breakdown between applications bridged by an AI agent that no single application owner ever authorized as its own risk surface, and most SaaS access reviews still examine one application at a time. The Cloud Security Alliance's State of SaaS Security 2025 report found 56% of organizations are already concerned about over-privileged API access across their SaaS-to-SaaS integrations.

Read the original at The Hacker News ↗