skills.rest
Mzane0803
Mzane0803COMMUNITY
Explore all skills by this author →
Website𝕏@MZane0803GitHub

scientific-schematics

Generate publication-quality scientific diagrams as PNG images via OpenRouter with iterative quality review.

Updated Aug 12, 2026
Search Tags:#png#image generation#flowchart#openrouter#publication figures#iterative refinement#scientific diagrams
One-click install
npx skills add https://github.com/Mzane0803/latent-minds-skills-marketplace --skill scientific-schematics-mzane0803
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: scientific-schematics
Source: https://github.com/Mzane0803/latent-minds-skills-marketplace/tree/main/plugins/research-writing/skills/scientific-schematics
Command: npx skills add https://github.com/Mzane0803/latent-minds-skills-marketplace --skill scientific-schematics-mzane0803

SYSTEM DOCUMENTATION & REQUIREMENTS

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

What problem does it solve? Creating publication-quality scientific diagrams normally requires manual drawing tools, templates, or coding, and iterating on quality is slow. This Skill generates diagrams from natural-language descriptions using the Nano Banana 2 image model and automatically reviews each result with Gemini 3.6 Flash against document-type quality thresholds, regenerating only when the score falls short. ## Core Features & Use Cases - Natural-language diagram generation: Describe flowcharts, neural network architectures, biological pathways, circuit diagrams, or system architectures and receive a PNG output. - Smart iterative refinement: Gemini 3.6 Flash scores each generation on scientific accuracy, clarity, labels, layout, and appearance; regeneration happens only below the threshold for the chosen document type (journal 8.5, poster 7.0, presentation 6.5, etc.). - Auditable review logs: Each run writes versioned images plus a JSON review log with scores, critiques, and early-stop reasons. - Use Case: A researcher needs a CONSORT participant flow diagram for a journal submission; they describe the screening, exclusion, and randomization counts and receive a threshold-checked figure with a review log. ## Quick Start Ask the assistant to generate a CONSORT flow diagram with 500 screened, 150 excluded, and 350 randomized participants saved to figures/consort.png for a journal document type.

Frequently Asked Questions about scientific-schematics

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

FAQPage Schema
How do I generate a scientific diagram from a text description?▼

Run scripts/generate_schematic.py with a natural-language prompt and an output path, for example a CONSORT flowchart description saved to figures/consort.png. The script generates the image with Nano Banana 2 and reviews it with Gemini 3.6 Flash automatically.

What diagram types can AI image generation create for papers?▼

The generator handles flowcharts such as CONSORT and PRISMA, neural network architectures, biological signaling pathways, circuit diagrams, system architectures, and block diagrams. Specific prompts naming components, counts, and flow direction produce the best results.

Does the schematic generator support vector output like SVG or PDF?▼

No, the pipeline outputs raster PNG only, with no vector path, DPI control, or CMYK color space. For journals requiring PDF, EPS, or 300 dpi TIFF, convert the PNG downstream and verify it at final print size.

Why does the review log show score null and reviewed false?▼

That means the Gemini review call failed due to a rate limit, content filter, or unparseable response, so no score was invented. The image is still generated and kept; inspect it yourself and re-run, since these failures are usually transient.

What API key and dependencies are required to run the generator?▼

You need an OpenRouter API key set as OPENROUTER_API_KEY, passed via --api-key, or stored in a .env file. The only Python dependency is the requests library, installable with uv pip install requests.

How many iterations and API calls does one diagram generation cost?▼

Each iteration costs two API calls: one image generation and one vision review. Iteration is capped at 2, so a run costs between two and four calls, and early stopping ends the loop once the score meets the document-type threshold.

Related Skills

Explore More Education & Research Skills→
lijiandaolijiandao
4

nature-writing

Draft Nature-style manuscript sections from claims, figures, notes, or Chinese drafts.

Community
Advanced
dralkhdralkh
74

venue-templates

Provide venue-specific academic formatting guidance and templates for manuscripts, papers, posters, and proposals.

Community
Advanced
pur3v4d3rpur3v4d3r
6

scientific-visualization

Create publication-ready figures with Matplotlib, Seaborn, and Plotly.

Community
Advanced
Zehong-WangZehong-Wang

scientific-schematics

Generate publication-quality scientific diagrams in SVG, EPS, and PDF formats.

Community
Advanced