According to the GATK, the --knownSites parameter accepts data in many formats:

--knownSites / -knownSites

A database of known polymorphic sites This algorithm treats every reference mismatch as an indication of error. However, real genetic variation is expected to mismatch the reference, so it is critical that a database of known polymorphic sites (e.g. dbSNP) is given to the tool in order to mask out those sites.

This argument supports reference-ordered data (ROD) files in the following formats: BCF2, BEAGLE, BED, BEDTABLE, EXAMPLEBINARY, RAWHAPMAP, REFSEQ, SAMPILEUP, SAMREAD, TABLE, VCF, VCF3

Source: https://software.broadinstitute.org/gatk/documentation/tooldocs/current/org_broadinstitute_gatk_tools_walkers_bqsr_BaseRecalibrator.php#--knownSites

Based on the data you've got, I would either aim to derive a VCF of the ~100 SNPs and then supply that to --knownSites, or else generate a SAMtools pileup. I would also try SAMtools mpileup, which allows you to generate a single pileup from multiple samples. For both of these ideas that I mention, you will of course have to first align your FASTA sequences to a reference bacterial genome (if available?).