I have no idea how UCSC handle our data. At a guess, I would say that these are Ensembl genes, Not sure, Ensembl proteins, Not sure and Links between Ensembl IDs and HGNC gene names respectively.

Vega is a browser of the manually curated Havana gene set. Ensembl also perform automatic annotation of genes using protein and nucleotide sequence databases, such as EMBL, Uniprot and RefSeq. Ensembl use the GENCODE gene set, which is made up of the Havana and Ensembl automatic gene set. Genes within the GENCODE set are labelled as being either Ensembl (from the automatic annotation), Havana (from the manual annotation) or merged (exact match between the automatic and manual annotation).