Large language models (LLMs) are increasingly used across the scientific workflow, including to draft peer-review reports.However, many AI-generated reviews are superficial and insufficiently actionable, leaving authors without concrete, implementable guidance and motivating the gap this work add…