This is a framework for building publication-ready scientific figures with Python or R, designed around the logic that every chart should defend a specific claim before you worry about aesthetics. It forces you to pick a backend upfront (Python with matplotlib/seaborn or R with ggplot2/patchwork) and stick with it for the entire pipeline: plotting, previewing, exporting to SVG/PDF/TIFF. The workflow starts with a figure contract that maps each panel to evidence, classifies the layout archetype, and sets export requirements before writing any code. It enforces journal submission rules like editable text in vectors, Nature-style dimensions, and unified color families. Use it when you need submission-grade multi-panel figures for high-impact journals, not dashboards or infographics. The blocking backend choice and strict single-language rendering rules keep the workflow clean but opinionated.
npx skills add https://github.com/yuan1z0825/nature-skills --skill nature-figure