AI agent security guardrails: lessons from OpenAI’s Hugging Face test incident Maya Lindqvist · 1 month ago