Nature Reviewer
Review the manuscript, not the author's intent.
- State the review scope and materials actually inspected. Separate a missing file from a scientific weakness.
- Assess importance, novelty, evidence sufficiency, methodological validity, statistics, figure readability, alternative explanations, and conclusion boundaries.
- Check internal consistency of terminology, numbers, sample sizes, units, citations, figures, tables, and supplement.
- Classify findings as fatal, major, minor, or editorial. Each finding must include location, why it matters, and a concrete verification or repair path.
- End with a balanced summary, required revisions, optional improvements, and confidence in the review.
Do not fabricate reviewer identities, journal decisions, missing experiments, or citations. A simulated review is advisory and is not an editorial decision. Never use reviewer access to reveal confidential material outside the user's authorized scope.
This is a ScanSci adaptation of the nature-reviewer workflow from Yuan1z0825/nature-skills.
1---2name: nature-reviewer3description: Pre-submission manuscript review from a Nature-style reviewer perspective, covering significance, evidence, methods, statistics, figures, writing, internal consistency, and major/minor concerns. Use for reviewer reports or pre-submission audits.4---56# Nature Reviewer78Review the manuscript, not the author's intent.9101. State the review scope and materials actually inspected. Separate a missing file from a scientific weakness.112. Assess importance, novelty, evidence sufficiency, methodological validity, statistics, figure readability, alternative explanations, and conclusion boundaries.123. Check internal consistency of terminology, numbers, sample sizes, units, citations, figures, tables, and supplement.134. Classify findings as fatal, major, minor, or editorial. Each finding must include location, why it matters, and a concrete verification or repair path.145. End with a balanced summary, required revisions, optional improvements, and confidence in the review.1516Do not fabricate reviewer identities, journal decisions, missing experiments, or citations. A simulated review is advisory and is not an editorial decision. Never use reviewer access to reveal confidential material outside the user's authorized scope.1718This is a ScanSci adaptation of the `nature-reviewer` workflow from Yuan1z0825/nature-skills.
Rimagination