| name | scholar-evaluation |
| description | Score and assess CS papers, proposals, literature reviews, and benchmark plans using a structured evaluation rubric. Use when a quantitative or semi-structured assessment is needed for research quality, novelty, technical soundness, empirical rigor, reproducibility, clarity, and submission readiness. |
Scholar Evaluation
Use this skill when the user wants more structure than a free-form review but does not need a full venue-specific peer review.
Good Fits
- comparing multiple candidate papers
- grading a draft literature review
- evaluating a research proposal before implementation
- checking whether a benchmark plan is rigorous enough
- estimating whether a draft is close to submission quality
CS-Oriented Dimensions
Score or discuss these explicitly:
- problem importance and positioning
- novelty or distinct value
- technical soundness
- empirical rigor or evaluation adequacy
- reproducibility and artifact clarity
- writing clarity and organization
- risk areas that could block publication or adoption
Local Resources
references/evaluation_framework.md
scripts/calculate_scores.py
Use the script when a numeric roll-up is useful. Otherwise keep the evaluation semi-structured and explain the score rationale.
Working Rules
- Do not hide major flaws behind an average score.
- Separate missing evidence from poor writing.
- State what would most improve the work.
- Use scoring to support judgment, not replace it.
Default Output
When no format is requested, return:
- dimension-by-dimension assessment
- strengths
- highest-impact weaknesses
- overall readiness judgment