scientific-schematics

Generate publication-quality scientific diagrams from natural-language descriptions.

2|Updated Apr 12, 2026
One-click install
npx skills add https://github.com/Clay-HHK/claude-config --skill scientific-schematics-clay-hhk
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: scientific-schematics
Source: https://github.com/Clay-HHK/claude-config/tree/main/skills/scientific-schematics
Command: npx skills add https://github.com/Clay-HHK/claude-config --skill scientific-schematics-clay-hhk

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires requests, python-dotenv, and includes scripts (resource) and references (resource) components.

What problem does it solve?

Convert informal natural-language diagram descriptions into publication-quality scientific figures without manual drawing, templates, or design work, saving researchers and authors time and ensuring consistent, accessible visuals.

Core Features & Use Cases

  • AI-powered generation: Produces high-quality diagrams (flowcharts, neural network architectures, circuits, biological pathways, system diagrams) from plain-language prompts.
  • Smart iterative refinement: Uses Nano Banana 2 for generation and Gemini 3.1 Pro Preview for quality review, repeating only when scores fall below document-type thresholds.
  • Document-aware quality: Applies different quality thresholds for journals, conferences, posters, presentations, theses, and grants and outputs versioned images and JSON review logs for reproducibility.

Quick Start

Generate a CONSORT participant flowchart for n=500 screened, n=150 excluded, n=350 randomized and save the final figure to figures/consort.png.

Frequently Asked Questions about scientific-schematics

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I generate publication-ready scientific diagrams from text descriptions?

To generate publication-ready scientific diagrams, provide a natural-language description of your figure. The tool uses Nano Banana 2 to create the diagram and applies document-type quality thresholds to ensure the output meets publication standards.

What types of scientific diagrams can I create using natural language prompts?

You can create neural network architectures, system and data-flow diagrams, biological pathways, circuit schematics, and flowcharts. These outputs are suitable for posters, conference papers, theses, and grant proposals.

Does the quality review process automatically refine scientific diagrams?

Yes, the quality review process uses Gemini 3.1 Pro Preview to evaluate diagrams. It applies up to two refinement iterations automatically, but only when scores fall below the specific thresholds set for your target document type.

Can I use this for generating CONSORT participant flowcharts for clinical trials?

Yes, you can generate CONSORT participant flowcharts by providing screening and randomization numbers in a prompt. It outputs the final figure as an image file and includes a JSON review log for reproducibility.

What are the limitations of iterative refinement for AI-generated scientific figures?

The iterative refinement process is limited to a maximum of two iterations. If a diagram still falls below the required document-type quality thresholds after these refinements, no further automatic corrections are applied.

How do I ensure reproducibility when generating technical schematics for research papers?

To ensure reproducibility when generating technical schematics, the tool outputs versioned images alongside JSON review logs. These logs document the quality review scores and refinement iterations applied during generation.