bracket-predictor

Quick Review

The skill provides a basic framework for bracket prediction with clear output formatting, but lacks substantive domain knowledge. While the description mentions key concepts (upset potential, chalk vs contrarian, historical trends), the instructions offer only generic guidance without concrete methodologies, data sources, statistical models, or bracket-specific strategies. The structure is clean but extremely shallow—no actual prediction algorithms, historical data references, or tournament-specific heuristics are provided. Novelty is low because a CLI agent could produce similar generic bracket analysis without this skill; the skill doesn't meaningfully reduce token costs or provide specialized knowledge that would be expensive to generate ad-hoc. To improve: add concrete upset prediction criteria (e.g., seed differentials, tempo matchups), historical tournament statistics, specific analytical frameworks (KenPom, NET rankings), and detailed examples of bracket picks with reasoning.