What problem does it solve? Evaluating mathematical modeling competition papers against national award standards is slow, inconsistent, and prone to missing format errors or undervaluing unconventional innovation. This Skill applies the CUMCM national award rubric to a paper and produces a structured review with per-dimension scores, located format issues, and prioritized revision suggestions. ## Core Features & Use Cases - Format compliance audit: Checks six dimensions (headings, figure/table numbering, formulas, references, abstract, page layout) and logs each issue with location, severity, and fix suggestion. - Five-dimension content scoring: Scores assumption validity, modeling creativity, result clarity, formatting, and references (0-20 each, totaled to 100) with reasons and confidence levels. - Paper-type-aware scoring: Classifies papers into traditional modeling, data analysis, machine learning, or engineering/signal types via a decision tree, then adjusts scoring so ML papers are not penalized for few figures and engineering papers not penalized for many figures or few references. - Innovation recognition with confidence and human-review triggers: Uses a five-type innovation taxonomy and seven-signal checklist to detect novel algorithms or cross-domain fusion, flagging low-confidence cases for manual expert review instead of scoring them down. - Use Case: A coach drops a student's 40-page PDF draft before submission; the Skill detects the paper type, flags a missing figure reference and an over-length abstract, scores creativity at the national-second-tier band, and returns a P0/P1/P2 revision list. ## Quick Start Send your modeling paper file to the agent and ask it to review this mathematical modeling paper against national award criteria and output the scored report with revision suggestions.
