Has anyone figured out how to use something akin to a land use raster as a covariate/predictor with openSTARS?

[KGBrennan]

I have been using openSTARS for the past few years and have always had to convert land use rasters to vector objects before building the stream network and making attribute calculations. Has anyone been able to use a land use raster or some other coded class type .tif in this package as a predictor covariate? I have followed the instructions and even broken rasters into subsets with a binary classification 0 and 1, is this the only way? The LU .tif that is already loaded with the NC example data is the only raster I have seen work on, but it has some structure different from a given terra rast or raster type tif. Cheers!