Seems to be some unnecessary repetition in our build & publish workflow

[lemanie]

It seems like there is some unexpected repetition occurring as some driver_version x kernel_package combinations seem to have the script run against them multiple times.

This will need to be verified using the raw workflow logs as the search in the web UI seems to not find the repetitions. A possible suggested cause is https://gist.github.com/posener/92a55c4cd441fc5e5e85f27bca008721.

Acceptance Criteria:

• For every 1 combination of 1 falco_version x 1 kernel_package, only process the package (check if mirrored then either skip/build & publish) once
• Add a unit test to validate