In the limit, If something takes 5,000 watts, you could run it for a few minutes/day, with that same RTG, provided you had suitable energy storage.
Perhaps they could gather grains of material, and just sort them, one at a time, only keeping the iron rich material, or use a permanent magnet to gather ferrous material. You could sinter the grains together using a microwave or laser pulse.
The results don't need great quality, just enough tensile and compressive strength to be mechanically stable during additive or subtractive manufacture.
Lots of minds have been thinking about refining metals for a very long time, but they haven't been thinking about doing it on Mars, with limited power, and very far outside the box of normal constraints, like cost.
This is one time capitalism doesn't apply at all... and most solutions assume capitalist incentives and costs, instead of going back to first principles thinking.
But thinking about it now, I can't envision that we here would be able to come up with something sublime/novel that a huge army of really-smart people haven't already after spending decades thinking about it. Then again, we have a lot of smart people concentrated in this forum, so who knows if a weird/silly conversation triggered by IT-minded people, acts as a catalyst for the engineer-lurkers that see it.
And then you land your digger, at a cost of at least a couple 100 million, and you have a dumb robot that digs a small hole?
We should do industrial build up, but for me that starts with solving cheap transportation from earth to Mars first. Then you can hope to bring things like nuclear reactors, that can actually produce the power you need significant work.
Currently we are going for science reason and very few people are even working on going with humans or industrialisation.