OpenAI’s Rogue AI Agents Were Probing Hugging Face Two Months Before Hack

An independent researcher found the agents hijacked Hugging Face accounts and mapped the platform’s defenses as early as May 13—activity OpenAI’s own incident report never fully described.