Currently there are no developers working on the Nyquist plug-in interface,
I don’t suppose you are a highly skilled C++ programmer by any chance?
Label tracks can be exported as text files.
Text can be input into a Nyquist plug-in, either be pasting into a text box (the easier method) or by the plug-in reading a text file (trickier).
The code in these plug-ins is somewhat convoluted. I’ve got some cleaner “detect sounds” code if you are interested in pursuing this as a plug-in. In fact (rummaging around…) this code isn’t too bad: https://forum.audacityteam.org/t/updating-sound-silence-finder/27236/1
Nyquist does not have access to the Audacity clipboard, but it is possible to copy audio from one track to another by temporarily holding the data in the SCRATCH symbol. There’s an example of this here: Running a plug-in once only - #7 by steve
It would be relatively easy to adapt this code to copy a selection from one track, apply a fade-in fade-out to the copied data and paste it into the next selected track.