document-illustrator

name: document-illustrator description: > AI-powered document illustration generator that analyzes document content and creates professional images. Use this skill when users need to generate illustrations for documents, articles, or notes. Supports three styles: gradient-glass (tech/modern), ticket (infographic/data), and vector-illustration (storytelling/education). Triggers: "generate illustrations", "create images for document", "配图", "插图", "document images"

Document Illustrator

Generate professional illustrations for documents using AI. Claude analyzes the content, summarizes key points, and generates images via Gemini API.

Workflow

Step 1: Read and Analyze Document

Read the document with Read tool, then intelligently identify core themes and key points. No specific format required - works with Markdown, plain text, or any readable format.

Step 2: Gather User Preferences

Ask the user three questions using AskUserQuestion:

1. Aspect ratio: 16:9 (landscape) or 3:4 (portrait)
2. Cover image: Generate a cover image summarizing the entire document?
3. Number of images: How many content images? (recommended: 3-10)

Step 3: Summarize and Confirm

Based on the requested number, intelligently group document content into themes. Present the summary to the user:

Content Summary Complete

Cover Image (if selected):
"[Document Title]"
- Core concept 1
- Core concept 2

Content Images (N total):
1. [Theme 1 Title]
   Includes: point A, point B, point C

2. [Theme 2 Title]
   Includes: point D, point E

...

Confirm to start generating? (Y/N)

Step 4: Generate Images

After user confirmation, call the Python script for each image:

python3 scripts/generate_single_image.py \
  --title "Image Title" \
  --content "Content description..." \
  --style-file references/styles/gradient-glass.md \
  --output /path/to/images/illustration-01.png \
  --ratio 16:9 \
  --resolution 2K

For cover images, add --cover flag.

Output location: images/ folder in the document's directory.

Style Reference

Three styles available in references/styles/:

Image Specifications

Environment Requirements

Python dependencies:

pip install google-genai pillow python-dotenv

API key: Set GEMINI_API_KEY in environment or create .env file in skill directory.

Content Grouping Principles

When summarizing document content:

• Completeness: Include all important information
• Logical flow: Group related content together
• Balance: Similar information density per image
• User control: Show summary for user confirmation before generating