Source: https://github.com/aipoch/medical-research-skills

When to Use

You need to batch convert a folder of images into a single target format (e.g., WebP) for distribution.
You want to reduce file sizes via compression while keeping processing fully offline (no network calls).
You need a repeatable, scriptable pipeline for CI/local automation (e.g., preparing assets for a website/app).
You want to preserve the source directory structure in the output directory during conversion.
You need best-effort batch processing where individual file errors are reported but do not stop the entire run.

Key Features

Batch conversion of common image formats using Pillow.
Configurable output format (default: webp) and quality (default: 80).
Optional recursive traversal of subdirectories while preserving folder structure in output.
Overwrite policy control to prevent accidental replacement of existing outputs.
Summary reporting (counts, errors) printed to standard output.
Local-only operation: reads from a specified source directory and writes to a specified output directory.