paper-reviewlisted

# Paper Review Review the argument before reviewing the prose. Help the author and editor see whether the paper has a real spine, where it breaks, and what kind of work would make it stand. ## Inputs and boundary Use the manuscript, its journal review template if available, and any requested review criteria. Read the whole relevant manuscript, including tables, figures, appendices, and references when they bear on the conclusion. This skill can prepare a private review draft. If a publisher requires an attestation that no generative AI was used to produce the report, do not mark it as true after using this skill. Leave any submission choice to Shiva. ## Review workflow 1. State the strongest possible version of the paper's idea. 2. Identify the one issue that determines whether the paper can stand. Usually it sits between the construct named, the mechanism claimed, the evidence available, and the conclusion drawn. 3. Separate what the manuscript directly observes, estimates or proxies, and infers. Do this only where the distinction affects the result. 4. Decide whether the repair is bounded or foundational. 5. Recommend the decision and write the report. Do not begin with a scorecard, a long list of small faults, or generic praise. ## Write the report Use this movement when it fits the paper. ```text the paper's strongest idea -> the key distinction it has not respected -> the evidence boundary -> the consequence for the claim -> the honest route forward ``` Write