[mikecerisano]

Fresh exFAT copies can acquire a macOS-generated AppleDouble companion whose bytes differ from the source. BitMatch correctly refuses to overwrite it, but that can leave a metadata conflict even when the media matches.

Adds Advanced → Exclude AppleDouble companion files, off by default, on Mac, iPad and iPhone. The preview checks bounded, paired AppleDouble v2 containers and lists exclusions before Start; the engine rechecks that exact list before writing. Ordinary ._ files, truncated fixed-size metadata, data-fork entries and unknown containers remain selected.

The full source inventory stays separate from the selected copy/verification set. Queue settings and history retain the policy and exclusion rows; PDF/CSV/JSON reports describe the omissions. Selected files retain independent readback and truthful progress. ASC MHL inventories cover the selected files and never replace existing histories. A filtered transfer cannot earn a whole-card safe-to-erase verdict. Diagnostics add only a count, without filenames or paths.

Validated locally:

• 54 real-driver transfers across APFS/HFS+/exFAT, with two destinations, first/repeated copies, genuine metadata conflicts and filtered copies. Selected destination hashes and unchanged source contents/paths/mtime checked.
• Actual Standard → ASC MHL handoff, including exFAT rereads; repeated histories retained byte for byte.
• Engine: 121 XCTest cases (15 skipped), plus 132 Swift Testing cases; zero failures.
• Mac: 1,069 passed, 7 skipped, zero failures, including queued-policy isolation, cancellation and reports.
• iOS: 49 passed, zero failures on iPad Air 11-inch (M4), iOS 26.5 simulator. Mac/phone/narrow-iPad controls rendered and inspected.
• Mac/iOS builds and git diff --check pass. All source changes rebuilt by the final app test runs.

See the investigation and validation notes. One earlier disposable-image teardown was busy; it was detached normally and removed, and the complete engine suite then passed independently. This is filesystem-driver and seeded UI coverage, not a physical-device walkthrough or proof of the reporter's exact cause.

Refs #16. Keep the issue open pending diagnostics and a reporter retest. No merge or release is included.