Our use-case is that we want to ship a templated script as part of a Habitat library package. This package doesn’t run a service, however, we do need to be able to find the Python interpreter that our packaged script needs to run (the script is called by a separate service).
I touched on this same need too over here: Service paths/config without service?
I ran into a similar scenario before. I ended up having the script in the library package accept arguments for anything that needed template rendered data. Then the calling service would use templating to construct the values for those arguments.