Hello,
I am the developer of the "resurrected" hi2txt utility

And what is requested is indeed feasible, but with a bit of coding:
# get games with nvram for hi scores
list all games having a <mame>\nvram\<game> directory
# get games with hi file for hi scores
list all games having a <mame>\hi\<game>.hi file
concat the two list, removing duplicates
for each games
call hi2txt to extract hi scores as readable stuff (sort of CSV output)
#as 1st line of the output is always the main table header...
#as 1st column is always RANK, the hi score is always in the 2nd column
read the 2nd line and extract NAME column
# optional: skip unknown name to keep only the one from your friends, or to minimize the displayed result in order to be readable in one screen?
put name+game in a structure storing name as a key, and game as a value, where multiple values are allowed
dump the content of the structure in something readable by your favorite front-end
What should be a valid output at the end to be integrated into a front-end? A jpg/png image?
Regards,
GreatStone.