image

name	image
description	Compress and resize large images for Claude Code context. Use when encountering "image too large" errors or when user wants to process/compress images before analysis.

Image Compression Skill

This skill helps you work with large images by compressing and resizing them to fit within Claude's context limits.

Capabilities

• Compress images to reduce file size
• Resize images while maintaining aspect ratio
• Convert images to optimized formats
• Handle multiple image formats (PNG, JPG, WEBP, GIF, BMP, TIFF)
• Automatic quality adjustment to meet size targets
• Batch process multiple images

When to Use

Use this skill when:

• Encountering "image too large" errors
• User wants to analyze a large image
• User wants to compress images before uploading
• Image file size exceeds recommended limits (typically >5MB)

How It Works

This skill uses Python with the Pillow library to:

1. Open and analyze image files
2. Resize images to reasonable dimensions
3. Compress with quality optimization
4. Output optimized images for Claude analysis

Default Limits

• Maximum dimension: 2048 pixels (width or height)
• Target file size: 1MB (adjustable)
• Default quality: 85 (JPEG/WEBP)

Usage Examples

# Compress image with defaults (max 2048px, target 1MB)
python image_compressor.py photo.png

# Resize to specific max dimension
python image_compressor.py photo.png --max-size 1024

# Target specific file size (in KB)
python image_compressor.py photo.png --target-kb 500

# Convert format
python image_compressor.py photo.png --format webp

# Custom quality (1-100)
python image_compressor.py photo.png --quality 75

# Batch process directory
python image_compressor.py ./images/ --batch

# Show info only (no compression)
python image_compressor.py photo.png --info

Output

Compressed images are saved with _compressed suffix by default:

• photo.png → photo_compressed.jpg

Use --output to specify custom output path:

python image_compressor.py photo.png --output optimized.jpg

Handling Large Images Workflow

For images that are too large:

1. First run with --info to see image details
2. Compress with appropriate settings
3. Use the compressed image for Claude analysis

Setup

Make sure you have Python and Pillow installed:

pip install Pillow