It's a little contrived, but what we do is to utilize the '@Library' line at the start of some pipelines to point to a known 'dead' commit. It should load as a valid library as long as it has a 'vars' folder in there (but can otherwise be empty).

This of course presumes that you have the ability to maintain a 'feature branch' in the offending library repo, which might be a bridge too far for the admins.

I will throw out there that my preferred approach would be to work with the company/admins to improve the library for all. Perhaps only a small subset of the utilities are truly 'global' for implicit load and the rest can be broken out to one or more explicit load libraries, for instance.

I'll also say here on the topic of long load times: I did stumble upon a newbie mistake I made here where I was trying to cram too many different types of things in one repo. This resulted in me utilizing the convenient 'library is in subfolder' option in the library's configuration to (I thought) ignore the rest of the folders in the repo containing the library. Turns out in fact this 'subfolder' library configuration ended up cloning the whole repository every time it was loaded :(.