Be aware AI reviewers of indirect prompt injection attack. Scholars sneaking phrases into papers to fool it. https://www.theregister.com/2025/07/07/scholars_try_to_fool_llm_reviewers/