Based on the few information you provided, and from my understating of your problem, please find below some general consideration which hopefully may help you in setting-up your simulations.
Since you have successfully built a geometry using the atmloc.f (for simplicity I will call this geometry/input “GCR.inp”), the next step would be integrating your detector geometry into the the GCR.inp. The adaptation of the detector geometry into the GCR.inp will mainly depend from the final goal of your study, on which I cannot comment further.
With regard to the source term to use in the GCR simulation, there are many threads in this forum you may consult, in addition to Chapter 16 of the FLUKA manual. I link a few of them below.
In general, you set-up the GCR generation by using the SPECSOUR and the GCR-SPE cards and by using a series of *.spc files which provide the ion composition of the galactic flux from Z = 1 to Z = 28.