If you get the PATRIC family ID from pretty much anywhere on the PATRIC website, but I usually end up getting it from a feature page. In this example, my PATRIC family ID is PGF_07220970.
We use linux piping to echo that (note we use the p3 echo not the standard echo) to get the features in the family, but we only want the patric_ids. Then we get those as sequences.
FAMILY=PGF_07220970 p3-echo $FAMILY | p3-get-family-features --ftype global --attr patric_id | p3-get-feature-data --attr aa_sequence > $FAMILY.tsv
This gives you a three column tsv file with the PATRIC family ID, the PATRIC protein ID, and the protein sequence. You can convert that to a fasta file like this:
perl -F"\t" -lane 'next if (/feature.patric_id/); print ">$F [PATRIC FAMILY $F]\n$F"' PGF_07220970.tsv > PGF_07220970.faa