https://en.wikipedia.org/wiki/Boeing_X-37#Development
You can estimate engineering limits from the rocket equation, delta-v = effective exhaust velocity * ln(mass ratio). In orbit, it probably uses hypergolic propellants, likely N2O4 and one of the hydrazines. These have exhaust velocities of about 3,350 m/s in vacuum. So, with the orbiter weighing 4.99 tonnes loaded, you could guestimate figures from 2,300 m/s (50% fuel fraction, payload+orbiter = 2.50 tonnes) up to 7,600 m/s (90% fuel fraction, payload+orbiter = 500 kg, if that's possible or useful).
https://en.wikipedia.org/wiki/Tsiolkovsky_rocket_equation
https://en.wikipedia.org/wiki/Liquid_rocket_propellants#Bipr...
For comparison, the far larger (~100 tonne) space shuttle orbiter can do 300 m/s of delta-v, carrying about 20 tonnes of monomethylhydrazine + N2O4. (Note that's a factor of two lower than what the rocket equation allows; I don't know why that is).
https://en.wikipedia.org/wiki/Space_Shuttle_Orbital_Maneuver...
The problem with that theory is that being in a particular orbit locks you into a very limited choice of targets. The X-37B is unlikely to have much in the way of cross range capability, so any potential target would have to be within a narrow band of it's orbital track. Also suppose you want to hit a target and you just flew over it, you'll have to wait a full orbit before you can strike, which is likely to take a lot longer than launching an ICBM. It's even worse than that though because if your orbit is inclined to the equator (required if your target is not on the equator), the earth rotates under you so any target you fly over on this orbit will have moved and won't be under your orbital track on your next orbit. You might have to wait dozens of orbits before you fly over it again.
Unless you have hundreds of warheads in orbit, or are prepared to wait days between attacks, the orbital weapons platform concept is a non-starter.