command_timeout doesn't work?

Hi! Thanks for your action.
We are using it for about two years. From time to time we face an unexpected behaviour: scp action lasts more than hour to copy files when it's normally takes for about 2-3 minutes to do it. Size of our files is stable, it's about 350-400 Mb. This problem is always solved by rerun of the github workflow.
But the question is about command_timeout parameter. If it doesn't interrupt action after specified time, what it's intended for?

Start by reviewing the scp action's handling of the command_timeout parameter and reproducing the reported behavior in a GitHub workflow with a 350–400 MB copy. Compare a normal run with the runs that exceed an hour, then verify that the configured timeout interrupts the action as documented.

Written by the indexing model from the issue text.