A few things to keep in mind when "reading" the GenBank flat-file format is that it is a "human-readable" file format and that it is a format that needs to work for virus sequences and for human sequences, and it also needs to work for small non-coding RNA or full-length human chromosomes.
And another thing to keep in mind is that it was started when GenBank (and I should say here, that it is pretty much the same for DDBJ, and although different format, the same information for EMBL/ENA format as well, as the three databases share and exchange data daily, so all three have the same sequences and information (including the features) associated with each record.
vkkodali pointed to a great reference, and all I will do here is to point to the original documents:
The release notes that come with every release of GenBank:
https://ftp.ncbi.nlm.nih.gov/genbank/gbrel.txt
And the The DDBJ/ENA/GenBank Feature Table Definition
https://www.insdc.org/files/feature_table.html
Which explains every feature type with excruciating details.