| name | reviewer-simulator |
| description | Use when simulating a journal/ML-conference reviewer for a CFD-AI/SciML manuscript, producing strengths, major concerns, reproducibility concerns, required experiments, and decision risk. |
| version | 0.3.0 |
| author | CFD-AI Paper Skills maintainers |
| metadata | {"short-description":"Reviewer simulation for CFD-AI/SciML manuscripts"} |
Reviewer Simulator
Trigger
Use for reviewer-style critique, pre-submission risk pass, or “what will reviewers attack?” when a broader review than CFD-only checklist is needed.
Reviewer criteria
- objective and motivation clear
- well-placed in literature
- claims supported by evidence
- significant contribution
- reproducible methods
- limitations honest
- figures/tables support claims
Output schema
Summary
Strengths
Major concerns
Minor concerns
Reproducibility concerns
Questions for authors
Required experiments
Decision risk
low / medium / high
Most valuable revision under time constraint
Anti-patterns
- Writing a friendly summary instead of a review with decision risk.
- Flagging generic ML weaknesses while missing CFD reproducibility details.
- Inventing reviewer concerns not grounded in the manuscript.
- Treating missing experiments as minor wording issues.
Verification
- Quote evidence where available.
- Mark inference vs manuscript-proven fact.
- Do not invent details.
