Figure Generation Skill for Claude
Generate publication-quality scientific figures using matplotlib/seaborn with a three-phase pipeline (query expansion, code generation with execution, VLM visual feedback). Handles bar charts, line plots, heatmaps, training curves, ablation plots, and more. Use when the user needs figures, plots, or visualizations for a paper.
Install
one command, takes just this skill from the repositorynpx skills add https://github.com/lingzhi227/agent-research-skills --skill figure-generationThe instruction itself
12 sections, as written by the authorScientific Figure Generation
Generate publication-quality figures for research papers.
Input
$0— Description of the desired figure$1— (Optional) Path to data file (CSV, JSON, NPY, PKL) or results directory
Scripts
Generate figure template
python ~/.claude/skills/figure-generation/scripts/figure_template.py --type bar --output figure_script.py --name comparison
python ~/.claude/skills/figure-generation/scripts/figure_template.py --list-types
Available types: bar, training-curve, heatmap, ablation, line, scatter, radar, violin, tsne, attention
Three-Phase Pipeline (from MatPlotAgent)
Phase 1: Query Expansion
Expand the user's figure description into step-by-step coding specifications using the prompts in references/figure-prompts.md. Determine: figure type, data mapping (x/y/color/hue), style requirements, paper conventions.
Phase 2: Code Generation with Execution Loop (up to 4 retries)
- Generate a self-contained Python script using the template from
scripts/figure_template.pyas a starting point - Write script to a temp file and execute:
python figure_script.py - If error: capture traceback, feed back, regenerate (see ERROR_PROMPT in references)
- If no
.pngproduced: add explicit save instruction, retry - On success: report the generated figure path
Phase 3: Visual Refinement
Read the generated PNG file and visually inspect using the VLM feedback prompts from references/figure-prompts.md:
- Does the figure type match the request?
- Are labels, titles, and legends correct?
- Is the color scheme appropriate and consistent?
- Are axis scales sensible? Is text readable at publication size?
If improvements needed: generate corrective instructions and re-execute.
References
- All MatPlotAgent prompts:
~/.claude/skills/figure-generation/references/figure-prompts.md - Figure templates:
~/.claude/skills/figure-generation/scripts/figure_template.py
Output
Both PNG (preview, 300 DPI) and PDF (vector, for paper) formats. Plus the LaTeX include code:
\begin{figure}[t]
\centering
\includegraphics[width=\linewidth]{figures/figure_name.pdf}
\caption{Description. Best viewed in color.}
\label{fig:figure_name}
\end{figure}
Quality Requirements
- DPI ≥ 300, or vector PDF
- Colorblind-friendly palette (no red-green only)
- All text ≥ 8pt at print size
- Consistent styling across all paper figures
- No matplotlib default title — use LaTeX caption
Related Skills
- Upstream: data-analysis, experiment-code
- Downstream: paper-writing-section, paper-compilation, slide-generation
- See also: table-generation
Other skills for the same job
different authors, same section of the catalogue
Create safety-bounded draft structures and run local deterministic checks for clinical case, diagnostic, trial, safety, and aggregate research reports. Use only with synthetic, de-identified, or aggregate inputs and verified source-fact manifests; every output requires qualified review.
Build evidence-traceable market research reports and assumption-driven market sizing or forecast scenarios. Use for market definition, industry and customer evidence, competitive landscapes, TAM/SAM/SOM reconciliation, forecast sensitivity, and auditable report scaffolds.
Advanced content and topic research skill that analyzes trends across Google Analytics, Google Trends, Substack, Medium, Reddit, LinkedIn, X, blogs, podcasts, and YouTube to generate data-driven article outlines based on user intent analysis
Shopify store command center. Orders, inventory, fulfillment, analytics, and store health. Works with any Shopify store via Admin API.
Production incidents dashboard. Reads ECS health, Sentry errors, CI failures. Offers to dispatch fix agents for active fires.
YOLO mode. Spawns 4 parallel C-suite agents (CEO, CTO, CFO, COO). Each analyzes the business from their perspective using ALL available data. Produces unfiltered Hard Truths report. After user types YOLO, autonomously runs the business for a day using /loop.
Research computing toolkit for optoelectronic information science and engineering, MATLAB/Octave, Python scientific analysis, signal processing, image processing, statistics, simulation, optimization, publication figures, sensor/time-series data, citation lookup, and common scientific libraries. Use when the user asks for MATLAB code, scientific Python, data analysis, plots, simulations, formulas, statistics, machine learning, optical/physical/materials computation, or reproducible research workflows.
> Verify statistics and claims in blog posts by fetching cited source URLs and checking if the claimed data actually appears on the page. Extracts all load-bearing claims (statistics, product or policy claims, ranking and comparative claims, named sources), validates cited URLs before fetching, and scores match confidence (exact match 1.0, paraphrase 0.7-0.9, not found 0.0). Flags uncited claims as UNVERIFIED. Use when user says "fact check", "verify statistics", "check sources", "validate claims", "factcheck", "source verification".
How to use it
Take lingzhi227/figure-generation from the repository into ~/.claude/skills for personal
use, or into .claude/skills inside a project.
The agent identifies a skill by the name field in its header. Two skills with the
same name cannot sit side by side — one of them will be ignored.