How about trying something like this:

esearch -db assembly -query "Homo sapiens[ORGN] AND latest[SB]" | efetch -format docsum | xtract -pattern DocumentSummary -element AssemblyAccession SpeciesTaxid SpeciesName FtpPath_RefSeq | sed 's/,.*//' | sort -k 3,3 | tee downloaded_genomes.tsv | cut -f 4 | sed -e 's/$/\/*genomic.gbff.gz/' | wget -i /dev/stdin

It is pretty ugly and I know it used to be so easy.