It seems strange that the wasp would de-leg the spider when it could sting and paralyze it. I also wonder if a Pompilid has the mandibles to cut legs off.

behavior is characteristic of this tribe of wasps (Auplopodini). The genera Phanagenia and Auplopus are known to do this also. If you note how the wasp is carrying the prey item (straddling instead of dragging backward like many other Pepsines and almost all of the pompilines do) it would make sense for the wasp to remove the legs of the prey item to facilitate transportation to a nest site. There's where the energetic tradeoff is: straddle-running is presumably more energy efficient as long as you're not hindered and dragging prey backward is presumably not as efficient since you're hindered by more friction. But is the energy saved straddle-walking the prey more than the energy it takes to remove the legs of the prey item? I don't know how you would test this, but it seems like an interesting question if you're interested in spider wasps like I am.