Usually you would pass input files by adding them to your bpipe command, so
bpipe run pipeline.groovy file_one file_two ....
These can then be resolved inside the pipeline using the $input variable ($inputs to get them all, $input1 for the first, etc), or you can use a split pattern to parallelise over the files, see:
If it's not really an input file but some other configuration type file then passing it as a parameter makes sense. For that case I'd recommend a comma separated list:
bpipe run -p m=file_one,file_two,...
Inside the pipeline script you can parse this:
m_values = m.tokenize(",")
Hope this helps!
Simon