A brief dip in a cyanide solution would likely do the trick. Part of
your problem is that the rubber tends to attack those components of
the alloy that are more reactive, so the things you might clean it
with will result in some surface enrichment of the gold content, and
thus a different, darker, gold color even if any oxidation or other
such stuff is cleaned up. Cyanide also attacks the gold, so it would
remove a very tiny bit of the entire surface, restoring the color
underneath. If exposure is brief, it shouldn’t change the texture or
finish much at all.
But of course, there are the safety issues with using cyanides. A
gold plating solution, if cyanide based, would likely do the same as
a straight cyanide solution, when used without the electrical
connections.
An electrostrip, which might not even be cyanide based, might do it,
but would be more likely to change the finish some. Some of these
would tend to electropolish, others might give you a kind of
crystalline look, or leave it duller than you might wish. Depends on
the solution, and the alloy composition.
The easiest solution to these quandaries might be to find a
commercial electroplating company. Many of these companies routinely
use cyanide based plating solutions, and perhaps they’d be willing to
dip it in one of their solutions, or electroclean it for you.
For future molding tasks, if you wish to avoid these problems
altogether, switch to silicone based mold rubbers, either the
vulcanized ones, or the RTV liquids.
All of these avoid the discoloration, since they have no sulphur
compounds in them to react with the gold. You will want to either
clean your mold frames and plates and sprue/formers very very
carefully, or use different ones, as traces of standard rubbers left
on the frames and plates will interfere with proper curing of the
silicone rubbers. Don’t switch back and forth between types using
the same frames, or you’re going to have trouble. The silicone
rubbers tend to be slightly less stretchy and strong, but not by much
unless your models are very complex with lots of undercuts, hollow
sections, and the like, when the silicones can be harder to cut a
good mold. But in general, they cut easier, not resisting the blade
at all, and the molds are self lubrication/releasing, needing no
sprays or powders to release the waxes. So your waxes can come out
with the exact degree of finish as your model had. Many of the
silicones are also lower in shrinkage, some of the RTV liquids are
zero shrinkage, so your waxes come out closer in dimensions to the
original (though even with no mold shrink, the waxes can shrink, and
the castings too, so if you want exact size duplication, you’ll have
to take additional steps (such as high expansion crystobalite
investments used for some dental casting).
Hope that helps
Peter