An Agent Skill for image processing. Gives AI agents (Claude Code, Cursor, Gemini CLI, etc.) the ability to convert image formats, resize images, and remove backgrounds.
| Operation | Description |
|---|---|
| Convert | Convert between JPG, PNG, WebP, SVG, BMP, TIFF, GIF |
| Resize | Resize with automatic aspect ratio preservation |
| Matting | Remove background colors with adjustable tolerance and feathering |
All scripts output structured JSON, making them easy to chain together for multi-step workflows.
Clone into your project's skill directory:
git clone git@github.com:FWcloud916/skill-simple-image-tool.git image-processingPython 3 and uv. Each script declares its own dependencies via PEP 723 inline metadata, so uv run handles installation automatically.
# Convert format
uv run scripts/convert_image.py logo.png webp
# {"success": true, "output": "logo.webp"}
# Resize (auto height)
uv run scripts/resize_image.py photo.jpg 512 0
# {"success": true, "output": "photo_512x384.jpg", "width": 512, "height": 384}
# Remove white background
uv run scripts/image_matting.py logo.png "#FFFFFF" 30 5 logo_matted.png
# {"success": true, "output": "logo_matted.png"}Compatible AI agents automatically discover the SKILL.md file and load the instructions when a task involves image processing. The agent then runs the bundled Python scripts to perform the requested operations.
See SKILL.md for the full agent instructions and references/REFERENCE.md for the detailed API reference.
This skill is part of the Simple Image Tool VS Code extension, which provides the same capabilities through a graphical interface and a GitHub Copilot chat participant (@image-tool).
MIT