Image Compression

The problem

Images often need to be smaller before upload, sharing, or front-end delivery. Saving originals as-is increases storage and loading costs. Adjusting quality, dimensions, and format by hand is repetitive, and batch work can easily lose directory structure. image-compress turns this into a reusable compression flow: it can process a single image, a directory, or files matched by a glob pattern, then display a summary table for verification.

How it works

The skill prepares a Python Pillow-based compression script and runs it with the supplied arguments. Core capabilities include:

• Quality control: --quality / -q sets output quality from 1 to 100, defaulting to 85; lower values generally produce smaller files.
• Format conversion: supports jpeg, png, and webp output, defaulting to the input format.
• Dimension limits: --max-width and --max-height constrain pixel size while preserving aspect ratio.
• Batch handling: for directories or glob input, output defaults to ./compressed and preserves the input directory hierarchy.
• Lossless option: --lossless is available for PNG and WebP only.

Boundaries and notes

The skill relies on a Python environment with Pillow. On some Windows setups, python3 may be a Store stub, while python is the usable command. JPEG output converts RGBA images to RGB. EXIF metadata is preserved unless explicitly stripped, so publish-ready assets should be reviewed separately. Lossless compression does not apply to every input format.