The standalone restore script was uploaded to each remote with the result
discarded and nothing logged, so a failed second upload left the restore script
missing from the remote with no warning while the archive still recorded
'complete'. Capture the result: log success, and on failure set uploadFailed
(-> warning status) and log the reason.
Fixes#226
Claude-Session: https://claude.ai/code/session_01WbGBN9VyRK61zczYWcCQ2i