All HKT data of the files provided on the command-line are written in the L1A products. This is incorrect, the data with timestamps more than 5 minutes before the coverage-time-start or 5 minutes after the coverage-time-end should be rejected.
The number of leap seconds since 1958 is hard coded in the module lv0_io.py, but should be obtained dynamically from tai_to_utc.py. Then we only have to keep this module up-to-date (using the IERS Bulletin C).