i'm quite new to this forum so i'll apologize if i come off as too new but here's what i've found after a few days of researching this.
1) there are quite a few ways of doing this
2) finding the safest way may not be the easiest
3) the microwavegoldkiln people have a good idea, but aren't willing to acknowledge that there are other ways.
Search youtube for "microwave kiln" and you'll see people using crucibles of varying types lined with graphite and magnetite to smelt anything from glass to brass and silver. The combination of graphite and magnetite specifically because graphite reacts well at lower temperatures but doesn't get hot enough, while the magnetite works great at higher temperatures.
As for safety, I'm not sure about you but I'd like to have something that I can move about easily, especially if it is going into and out of a microwave. I like the idea that creek's acquaintance had as it makes the whole process easier.
As for the premade kit, they use a 2300F rated rigidized silica board. This is fine but I'd prefer something that provides a greater safety margin, and possibly some containment in case of failure. Also they use (most likely a graphite/magnetite blend?) additive in their custom flux and consider this to be the best option. However I've found an alternative that I really like in my search.
I've found this link from a government study that used a SiC crucible as a heating element, with another crucible inside it using a lab grade microwave.
http://www.osti.gov/scitech/servlets/purl/5671871
Admittedly the frequencies were tuned in this microwave and probably more directional but SiC is a good susceptor and with either more time, or a stronger microwave (and probably including rotation) this should be a good working 'blueprint'.
I've contacted a few insulation manufacturers to inquire about high density Alumina plates and tubes to use as an insulator for the whole process because they are microwave transparent AND stable to nearly 3600F, thought this is above the operational temperature of the SiC crucible it does give that safety factor I was looking for. It would also remove the need for custom flux.
As to any metals inside the crucible (which is inside the SiC crucible) sparking - it won't happen as the SiC absorbs almost 100% of the microwaves converting them to thermal energy.
So, basically an Alumina chamber with a SiC crucible that is slightly larger than your working crucible should work to (s)melt most things up to 1200C (@2200F) (without much risk of the SiC crucible failing) and up to 1600C(@2900F) for short periods, though the microwaves may severely affect it at higher temperatures.
I don't want to reinvent a wheel, I just don't have the desire to drag around 200lbs of gear to smelt my placer gold on site when 50lbs will do.
When I get everything together and test it I'll let you know how it goes!