Currently, there's little use for space tugs because:
- Propellant dominates the mass budget for sending satellites into orbit;
- The propellant needed to send a space tug plus a satellite to a given orbit, and then return the tug, is much more than that needed to send the satellite alone;
- All the propellant involved needs to be launched from Earth.
At present, the usual strategy for communications satellites in geosynchronous orbits is to carry enough propellant to keep station for their expected operational lifetime, then send them to a graveyard orbit and launch a replacement satellite that has much improved communications electronics, able to handle much more data per kilogram of satellite and solar panel. With that in mind, the rest of the hardware usually isn't built to last much longer than the on-station lifetime. Making longer-lasting satellites would increase costs. Making satellites that can have electronics modules swapped out for newer ones by a robotic tug to extend their service life would increase costs.
Even fully reusable launchers -- of which none exist yet -- will burn about 50 tons of propellant for every ton they take to LEO. A tug taking propellant to GEO and back to refuel satellites will, likewise, burn much more fuel itself than it will be able to deliver.
If there was an orbital station that was able to do a lot of repair and maintenance work on a large volume of satellites, I could -- just barely -- imagine a scenario where space tugs were useful, but I suspect that the physics and engineering of it would actually favor satellites using ion thrusters carrying enough extra propellant for them to return to the "garage" every few years on their own, rather than being towed there.