hierarchical-models

Quick Review

Strong skill providing practical hierarchical Bayesian modeling patterns with excellent task knowledge including Stan and JAGS implementations, centered vs non-centered parameterizations, and diagnostic guidance. The description adequately conveys the skill's scope (random effects, partial pooling, parameterization strategies). Structure is clear with logical sections covering when to use, core concepts, implementations, and diagnostics. Novelty is moderate: while hierarchical models require specialized knowledge, a capable CLI agent with access to Stan/JAGS documentation could construct these patterns with sufficient prompting, though this skill notably reduces token cost and provides decision heuristics (e.g., when to use non-centered). The skill would benefit from slightly more detail on invocation patterns (e.g., how to apply to new datasets) but overall provides actionable, well-organized Bayesian modeling guidance.