This is the text (as seen in notepad++) of MB 657.
I had to zip the .txt, just unzip it.

So at a minimum my unwritten Perl script (which works perfectly in my imagination btw) would at a minimum parse all the GGR files, no matter on what continent or on ships at sea, hunting for the one with the largest value in line 3. Then, it would go on to ascertain that all the colors were unique. Then it would output gold sovereigns. But not too many.

I created this post with the hope that a GIMP script might be written: I apologize that this is not about a script, per se.

However, having acquainted myself with the clear exposition of GGR by user jsbueno on stackoverflow, I've thrown together a Perl script to create a gradient of randomized colors which may become grist for someone else's script mill, to wit:
#!/usr/bin/perl

open FILE, ">New256color.ggr" or die $!;
print FILE "GIMP Gradient\n";
print FILE "Name: New256color\n";
print FILE "256\n";
for($seg=0;$seg<1;) {
$segm=$seg+1./512.;
$segr=$seg+1./256.;
print FILE "$seg $segm $segr ";
$red=rand();
$green=rand();
$blue=rand();
print FILE "$red $green $blue 1 ";
print FILE "$blue $red $green 1 "; # 2nd, blend of L and R colors across segment
#print FILE $red $green $blue 1 "; # Orig. pure color across segment
print FILE "0 0\n";
$seg=$segr;
} # for seg

close FILE or die $!;

Actually, there are more than 256 colors, I think. Originally, I used the same values of RGB for both left and right endpoints of each segment, but at the last moment, I got the bright idea to alter the right-hand endpoint in a trivial way. The result appears to be that GIMP creates a new mini- or micro-gradient on each of the 256 segments. However Colors->Info->Histogram reports pixel counts only for values 0-255, so although ANY collection of 256 colors is possible, only 256 are stored in a palette. (No doubt long-time users of GIMP are much more familiar with its constraints than I am.)

Here's a comparison between (a closeup of) the 2nd (blended segments, above) and original gradients: