Well, lets ponder this a little further. Currently there are no volcanoes in Yellowstone. So we know that the rock layer is sufficient. We also know that the rock keeps losing heat with time because it has higher than average temperature. But we also know that the rock keeps gaining heat from magma pushing up from some deep underground crack.
What we have to do is measure the net heat gain and calculate how much additional heat a geothermal plant has to extract from the rocks to turn it into net heat loss. Of course the plant will have to extract it from the hotter spots in order to prevent any local maximums from causing volcanoes. But that is the most efficient option for a geothermal plant anyways.
As far as measuring the net heat gain -- it should not be that hard. All you need is to sprinkle a couple of hundred sensors in the rock, monitor them for a while to be able to remove surface effects (like the sun and ambient air temperature) and you should be able to get a decent estimate. In fact I would not be surprised if seismologist already have a pretty good estimate of the heat gain.
Then, when you build the plant, you can be certain that as long as heat is being drained from the rock, magma will keep solidifying and the rock layer will keep getting thicker, etc.
Of course a good earthquake can screw up all of these plans, but an earthquake can destroy any power plant. But then again Yellowstone is one of the most earthquake prone areas of the world, so that perhaps is the biggest danger of this plan.