One should clone vdjdb-db repo and this repo to the same directory, say ~/vcs/ and then navigate to ~/vcs/vdjdb-db and run bash release.sh which will build the VDJdb database, navigate to motifs repository and run motif inference.
This folder lacks pools/ directory with control TCR sequences that is ~500Mb compressed, it will be automatically downloaded from this Zenodo record when running compute_vdjdb_motifs.Rmd or release.sh.