scientific-visualization

SUMMARY

What it does

This Agent Skill provides procedural guidance for creating and auditing scientific figures that are truthful, accessible, and publication-ready. It covers figure design principles, multi-panel layouts, honest encoding of uncertainty and missing data, color and contrast review, image metadata validation, and journal export planning. The skill includes bundled command-line tools for inspecting image metadata, auditing palette contrast, planning publisher exports, and previewing styles. It emphasizes preserving raw data, avoiding deceptive visual encodings, and verifying journal-specific requirements. The skill is designed for use with Matplotlib, Seaborn, or Plotly and requires Python 3.11+ and uv for pinned examples. It is distributed under the MIT license.