Use of aligners (preferably STAR) for read-barcode matching

My goal is to match a custom single cell type library with unknown barcode locations across reads but known sequences: I have made a ‘genome’ of ~70nt of each ‘read’ (17mX70nt contigs) and have aligned putative barcodes (~70kX32nt) against this. I specifically want the 32nt barcodes and 70nt reads to be locally aligned very similar to a standard Smith-Waterman alignment and mismatches on either end of the 32nt be penalized and not soft-clipped. Naturally the reads (barcodes here) need to be mapped to 100s or 1000s locations and soft-clipping is not desired here. These are my STAR options:

STAR 
--runThreadN 16 
--outFileNamePrefix my_run 
--alignEndsType EndToEnd 
--outSAMmode NoQS 
--scoreDelOpen 0 
--scoreDelBase -1 
--scoreInsOpen 0 
--scoreInsBase -1 
--outFilterMultimapNmax 10000 
--winAnchorMultimapNmax 10000 
--seedPerReadNmax 10000 
--seedPerWindowNmax 10000 
--outSAMattributes NH HI AS nM NM MD 
--outFilterMismatchNmax 2 
--genomeDir ./my_ref/ 
--readFilesIn my.fasta

This setting works relatively well however I would like to ask if there is there a set of aligner parameters to allow me to swap my reference and reads in order to have a more convectional mapping?