This is a solid Polars-based workflow for CSV, JSON, and Parquet analysis. It includes reference docs for loading, transformations, aggregations, time series, and visualization, plus ready-to-use exploration scripts. The iteration checkpoints are genuinely helpful: they prompt you to show the user data shape and samples after loading, summary stats during exploration, before/after comparisons during transformation, then findings and charts. The quick start example is clean, and the lazy evaluation guidance is the kind of performance tip that actually matters when you're working with files that don't fit comfortably in memory. If you do regular data wrangling, this gives you patterns that work without reinventing the wheel each time.
npx skills add https://github.com/artificialanalysis/stirrup --skill data_analysis