I was looking at the metadata for the ExperimentHub object:

> library(ExperimentHub)
> eh <- ExperimentHub()
> md <- mcols(eh)
> names(md)
 [1] "title"              "dataprovider"       "species"            "taxonomyid"         "genome"            
 [6] "description"        "coordinate_1_based" "maintainer"         "rdatadateadded"     "preparerclass"     
[11] "tags"               "rdataclass"         "rdatapath"          "sourceurl"          "sourcetype"

These are essentially the same as what is provided in the metadata.csv when submitting a new resource. However, I noticed that it's missing some columns like DispatchClass and Location_Prefix. Why are these hidden?

Stylistic really. No real reason but to try to limit to what we felt was more important to search. Generally source information is more ideally searchable and more meaningful than the actual location of the (often manipulated or curated) data download path; plus it was only recently we realized the popularity so when originally designed the data was hosted on Bioconductor servers so the location_prefix was mostly irrelevant; now that we are encouraging and favor hosting elsewhere it might be appropriate to reconsider but most information could be achieved with the already present source information.
Similarly DispatchClass is more internal or backend information on how the data should be loaded in R during the call to experimenthub to download the resource; The equivalent end user information is captured by rdataclass instead indicating what object you actually expect to get back in R with the call.