[Enhancement] Find duplicate files using size filters and streamed hashes #8

Priority: P2
Parent roadmap: #1

Goal

Help power users find exact duplicate files efficiently without becoming an unsafe bulk cleaner.

UX

• Select one or more Finder folders → Find Duplicates…; optional multi-folder comparison.
• Group by file size, then fast partial fingerprint if useful, then full streaming SHA-256 for exact byte equality. Do not use only filenames, modification dates or perceptual similarity to call files exact duplicates.
• Native results window grouping verified duplicates with file size, path, last modified date and quick reveal/open.
• Mark potential reclaimable size, clearly accounting for APFS clone/hardlink behavior where detectable.
• Optional manual move-to-Trash of explicitly selected duplicate copies with a review dialog. No automated permanent deletion and no default destructive action.
• Skip symlinks or handle explicitly; avoid traversing same directory via links; handle iCloud placeholders, access errors, bundles, hidden files and external volumes deliberately.

Acceptance criteria

• No duplicate false positives from truncated/partial hashes.
• Stream files in chunks; scanning large directories does not load entire files into memory.
• Cancel/progress and bounded concurrency; never block Finder.
• Preview/selection makes it obvious which path would be moved to Trash; never trash all copies in a group.
• File-identity recheck before trashing prevents deleting files changed since scan.

Tests

Identical-content/different-name files, equal-size/different-content, APFS clones/hardlinks, unicode/case differences, permissions, directory symlink cycles, iCloud/external-unmount, cancellation and race changes.