image-generation

Kjdragan's avatarfrom Kjdragan

AI-powered image generation and editing using Gemini. Use when Claude needs to: (1) Generate images from text descriptions, (2) Edit existing images with instructions, (3) Create infographics or charts, (4) Generate visual assets for reports/presentations, (5) Work with .png, .jpg, .jpeg, .webp files for editing.

⭐0stars🔀0forks📁View on GitHub🕐Updated Jan 4, 2026
[Visual Content Workflows]

When & Why to Use This Skill

This Claude skill empowers users with AI-driven image generation and sophisticated editing capabilities directly within their workspace. By leveraging Gemini's advanced vision models, it solves the challenge of creating high-quality visual content from text descriptions and streamlines the process of modifying existing images. It is an essential tool for professionals needing to produce infographics, charts, and custom visual assets for reports and presentations without leaving the AI chat interface.

Use Cases

  • Automated Infographic Generation: Transform complex data and statistics into professional, minimalist infographics and charts for business reports.
  • Instruction-Based Image Editing: Modify existing photos or graphics by adding elements, changing backgrounds, or adjusting styles using natural language commands.
  • Presentation Asset Creation: Generate custom high-resolution backgrounds and thematic illustrations for PowerPoint (pptx) slides to enhance visual storytelling.
  • Document Enrichment: Automatically create and embed relevant diagrams or figures into technical documentation and HTML reports to improve readability.
  • Rapid Visual Prototyping: Quickly iterate on creative concepts, brand assets, and UI mockups during the design and brainstorming phase.
nameimage-generation
description"AI-powered image generation and editing using Gemini. Use when Claude needs to: (1) Generate images from text descriptions, (2) Edit existing images with instructions, (3) Create infographics or charts, (4) Generate visual assets for reports/presentations, (5) Work with .png, .jpg, .jpeg, .webp files for editing."

Image Generation & Editing

Overview

Generate and edit images using Gemini 2.5 Flash Image model via MCP tools.

Quick Start

Text-to-Image Generation

result = generate_image(
    prompt="A modern infographic showing renewable energy statistics",
    output_dir="/path/to/work_products/media"
)

Image Editing

result = generate_image(
    prompt="Change the background to a sunset over mountains",
    input_image_path="/path/to/original.png",
    output_dir="/path/to/work_products/media"
)

Preview Generated Image

# Auto-launch viewer after generation
result = generate_image(
    prompt="Blue square with rounded corners",
    preview=True  # Opens Gradio viewer
)

# Or preview any existing image
preview_image("/path/to/image.png")

Available Tools

generate_image

Primary tool for image generation and editing.

Parameters:

  • prompt (required): Text description or edit instruction
  • input_image_path (optional): Path to source image for editing
  • output_dir (optional): Output directory (defaults to work_products/media/)
  • output_filename (optional): Custom filename (auto-generated if not provided)
  • preview (optional): Launch Gradio viewer after generation
  • model_name (optional): Gemini model to use. Defaults to gemini-3-pro-image-preview.
    • Options: gemini-3-pro-image-preview, gemini-2.5-flash-image.

Returns: JSON with output_path, description, size_bytes, and optionally viewer_url

describe_image

Get a short description of an image for generating meaningful filenames.

Parameters:

  • image_path (required): Path to image file
  • max_words (optional): Maximum words in description (default 10)

preview_image

Launch Gradio viewer to display an image.

Parameters:

  • image_path (required): Path to image file
  • port (optional): Port for Gradio server (default 7860)

Best Practices

Prompt Crafting

  • Be specific: "modern, minimalist infographic with blue gradient"
  • Include style: "photorealistic", "illustration", "line art", "cartoon"
  • For charts/infographics: Include data points in the prompt
  • For editing: Describe what to change AND what to preserve

Examples

Infographic:

generate_image(
    prompt="Infographic showing 3 key statistics: 45% growth, $2.3B market, 120 companies. Modern flat design with blue and green color scheme."
)

Chart:

generate_image(
    prompt="Clean bar chart comparing renewable energy adoption: Solar 35%, Wind 28%, Hydro 22%. Use professional colors, include labels."
)

Editing:

generate_image(
    prompt="Add snow to the mountain peaks, keep the sunset colors",
    input_image_path="work_products/media/sunset_mountains.png"
)

Integration with Other Skills

Reports (report-creation-expert)

Generate custom graphics for HTML reports:

# Generate infographic
img_result = generate_image(
    prompt="Infographic about AI market trends",
    output_dir="work_products/media"
)

# Embed in HTML report
html_content = f'''
<img src="file:///{img_result['output_path']}" alt="AI Trends" style="max-width: 100%;">
'''

PowerPoint (pptx skill)

Create custom slide backgrounds and visual elements:

  • Generate images first, then reference paths in PptxGenJS
  • Use addImage() to embed generated graphics
  • Create infographics for data-heavy slides

Documents (docx skill)

Generate embedded figures and diagrams for Word documents.

Troubleshooting

Error: "GEMINI_API_KEY not set"

Ensure GEMINI_API_KEY is set in .env file.

Error: "No image data in response"

  • Check your prompt is clear and specific
  • Ensure Gemini 2.0 Flash model supports the request
  • Try simplifying the prompt

Preview not working

  • Gradio viewer script must exist at .claude/skills/image-generation/scripts/gradio_viewer.py
  • Port 7860 must be available
  • Run manually: python scripts/gradio_viewer.py /path/to/image.png

Output Conventions

  • Directory: {workspace}/work_products/media/
  • Filename pattern: {description}_{timestamp}.png
  • Format: PNG (lossless)
  • Description: Auto-generated from image analysis

Dependencies

  • google-genai>=1.56.0 - Gemini API client
  • gradio>=6.2.0 - Optional viewer UI
  • pillow - Image processing