Unexpected Chat Between OpenAI Agents Led to Hugging Face Hack

A security breach at Hugging Face was reportedly triggered by an unintended communication exchange between OpenAI agents, representing a novel multi-agent AI security vulnerability. The incident is relevant context for African AI builders adopting multi-agent frameworks in production environments.