matplotlib
Low-level plotting library for full customization — use when you need fine-grained control over every plot element, creating novel plot types, or integrating with specific scientific workflows for PNG/PDF/SVG export.
Fine-grained control over every element of a Python plot
Trigger phrases
Phrases that activate this skill when typed to Claude Code:
plot this with matplotlibcreate a custom figurematplotlib axesexport plot to PDFannotate this chart
What it does
matplotlib is a Claude Code skill from K-Dense AI’s scientific-agent-skills repo. It turns Claude into a matplotlib expert covering both the pyplot interface (MATLAB-style) and the object-oriented API (Figure/Axes) — with guidance on the full matplotlib hierarchy, customization of every visual element, multi-panel layouts, and export to PNG, PDF, SVG, and EPS.
A session produces working Python code for the requested visualization — from simple line plots to complex multi-axis figures with insets, custom tick formatters, log scales, colorbars, and annotation arrows.
When to use it
Reach for it when:
- You need precise control over a specific plot element that higher-level libraries (seaborn, plotly) abstract away
- You’re creating a novel plot type not natively supported by seaborn or other wrappers
- You need export to a specific format (EPS for journal submission, SVG for vector editing) at a specific DPI
When not to reach for it:
- Quick statistical plots with sensible defaults — use
seaborn - Publication-ready multi-panel figures with journal styling — use
scientific-visualization - Interactive plots for web deployment — consider plotly
Install
Copy the SKILL.md from K-Dense AI’s matplotlib folder into .claude/skills/matplotlib/ in your project.
Trigger phrases: “plot this with matplotlib”, “create a custom figure”, “matplotlib axes”, “export plot to PDF”.
What a session looks like
A typical session has three phases:
- Plot specification. Describe the data and the desired visualization — plot type, axes, number of panels, and any specific styling or annotation requirements.
- Code generation. Claude generates matplotlib code using the object-oriented API (Figure/Axes) for flexibility, with explicit setup of figure size, DPI, font properties, and color choices.
- Refinement. You run the code, describe any adjustments (axis limits, label positioning, legend placement), and Claude makes targeted edits to specific Artist objects rather than regenerating the full figure.
Receipts
Where it works well:
- Custom annotation layers on top of existing plots — arrows, text boxes, span overlays, bracket markers — where Claude knows every relevant Artist method
- Precise axis control for log-scale plots, dual y-axes, and broken axis figures that break seaborn’s abstractions
Where it backfires:
- Complex interactive plots — matplotlib’s animation and widget interfaces require more back-and-forth than plotly’s declarative model
- First-pass figure aesthetics for non-technical audiences; matplotlib defaults are functional but not polished without explicit styling
Pattern that works: always use the object-oriented API (fig, ax = plt.subplots()) rather than pyplot state machine; it makes every subsequent edit precise and avoids the global state bugs that make matplotlib debugging frustrating.
Source and attribution
Originally authored by K-Dense Inc.. The canonical SKILL.md lives in the matplotlib folder of their public scientific-agent-skills repository.
License: MIT. Install, adapt, and redistribute with attribution preserved.
This page documents the skill from a practitioner’s perspective. For the formal spec and any updates, defer to the source repo.