I believe the .raw files are not necessary for building a spectra library from MaxQuant results. The spectra are actually stored in the msms.txt file.
By the way, this page has the list of which files are necessary for building spectral libraries from various types of peptide search results:
I know that .raw files are necessary for extracting chromatograms in Skyline, and also when building libraries from other types of peptide search results it it is necessary for Skyline to have access to the .raw files or the .mzML files.
Spectral libraries are used for two different purposes in Skyline. One is like a traditional spectral library where Skyline looks at the spectra in order to decide which transitions are most likely to be present so that you can decide which chromatograms to extract in an SRM, DIA, or PRM experiment. For these sorts of spectral libraries you would create them from one set of data, and then use the library to design future experiments.
The other thing that you can use spectral libraries for is as a way of converting peptide search results into something that Skyline understands, but all that Skyline is actually interested in are the retention times of the peptides in the particular .raw files that were searched.
You might find it helpful to look at the MS1 Full Scan filtering tutorial:
I believe that tutorial shows you how to create a spectral library from DDA search results and use those results to decide what MS1 chromatograms to extract and at which retention times to find the peptides.