How does the smart iteration feature save costs?

The skill evaluates the first generation against a quality threshold (e.g., 8.0/10 for architecture). If the score is high enough, it skips the second iteration, reducing API calls and processing time.

Does this require an external API key?

Yes, you must provide a GEMINI_API_KEY from Google AI Studio to power the image generation and the AI quality review process.

What types of diagrams can I generate with this skill?

You can generate a wide range of technical visualizations including system architectures, flowcharts, ERD data models, sequence diagrams, C4 models, and CI/CD pipeline diagrams.

Can I modify a diagram after it has been generated?

Yes, you can use the /nano-banana:edit command to provide natural language instructions for modifying an existing PNG file, such as adding a new service or changing a relationship.

What are the document type thresholds?

The skill uses different quality bars depending on the context; for example, 'specification' and 'journal' require high scores (8.5), while 'presentation' or 'readme' have lower thresholds (6.5-7.0) for faster results.

AI Diagram Generator (Nano Banana)

Name: AI Diagram Generator (Nano Banana)
Author: flight505

byflight505

•

Design Tools

Generates publication-quality technical diagrams and system architectures using AI-powered quality review and iterative refinement.

The Nano Banana Diagram skill transforms natural language descriptions into professional, high-resolution technical visualizations ranging from microservice architectures to complex sequence diagrams. By leveraging Gemini 3.1 Pro for both generation and automated quality auditing, it ensures that every output meets specific professional standards—such as colorblind-friendly palettes and high readability—while optimizing API usage through smart iteration logic that only regenerates when quality thresholds aren't met. It is ideal for developers and architects who need to maintain up-to-date documentation, READMEs, or presentation materials directly within their development workflow.

Key Features

01Publication-Ready Output: Delivers high-contrast, 4K-ready visualizations with readable fonts and colorblind-friendly designs.

02Intelligent Editing: Modify existing diagrams using natural language commands to add layers or change components without full regeneration.

03Smart Iteration Logic: Automatically regenerates diagrams only when quality scores fall below document-specific thresholds to save API costs.

04Comprehensive Diagram Support: Creates architecture diagrams, flowcharts, ERDs, sequence diagrams, C4 models, and data pipelines.

053 GitHub stars

06AI-Powered Quality Review: Gemini 3.1 Pro audits every diagram against five professional criteria including accuracy and readability.

Use Cases

01Creating presentation-ready system designs and C4 models for stakeholders, business proposals, or academic publications.

02Generating sequence diagrams and data flowcharts for PRDs, engineering specifications, and API documentation.

03Visualizing complex microservices architectures and cloud infrastructure for technical documentation and READMEs.

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add flight505/nano-banana diagram

For use in Claude.ai and ChatGPT

Download Skill