Tests multiple download methods with a known 1KB binary file. Each file should start with bytes: 50 4b 03 04 (PK header). After downloading, check with: xxd file.xls | head -3
If corrupted, you will see ef bf bd (UTF-8 replacement chars) in the hex dump.
After downloading, run on the worker:
for f in ~/Downloads/*.xls ~/Downloads/*.xlsx ~/Downloads/*.bin; do [ -f "$f" ] || continue echo "=== $(basename $f) ===" xxd "$f" | head -2 echo "efbfbd count: $(xxd "$f" | grep -c 'efbf bd')" echo "size: $(wc -c < "$f")" echo done
Expected clean first line: 504b 0304 0405 0607 0809 0a0b 0c0d 0e0f
All files should be exactly 1024 bytes and have 0 efbfbd matches (except test 6 which is intentionally corrupted).