Image Auto Classification and Archiving

Problem

Image files often accumulate across directories. Manual archiving by file name, capture time, or use case can miss items, overwrite names, and leave no audit trail. Mixed screenshots, assets, and photos in bulk JPG/HEIC folders are expensive to categorize and hard to explain after the fact.

How It Works

The skill turns archiving into a verifiable workflow:
- Recursive scan: recognizes jpg, png, gif, bmp, webp, heic, and heif, flagging abnormal files.
- Dual classification: date-year-month and date-year-month-day prefer EXIF capture dates, falling back to file timestamps; type mode uses file-name keywords, EXIF, and size heuristics.
- Safe operations: supports move or copy; duplicate names get suffixes automatically, and move records after transfer so unfinished items do not affect completed ones.
- Structured report: outputs a Markdown report with an operation summary, item-level details, and exception notes, suitable for logs or archives.

Boundaries and Notes

It is suited to local image-folder cleanup, not database-style asset management. Files above --max-size-mb are still processed but may rely only on EXIF and keywords, skipping pixel-level checks. It depends on Pillow and optional pillow-heif; missing HEIF support limits related capabilities. Preview with --dry-run first, then confirm source, mode, and operation.