On-orbit repairs are not being done these days. There are no spacecraft from which repairs can be done on orbit.
The Shuttle era showed on-orbit repairs to be very expensive (on the order of the cost of a new satellite), making it not worth bothering for a commercial satellite. After several experiments early on in the Shuttle program, repair missions dwindled until only the Hubble service missions were left.
What would it take to run a repair mission?
There are two major factors you need to make a repair mission possible:
The Shuttle showed that on-orbit repairs become much easier if you can attach the satellite to the spacecraft, and if the spacecraft has a manipulator arm. The arm can be used to move large components around, and it can provide astronauts with a surface to stand on/be attached to.
The satellite also has to be designed for it, with a docking interface, handrails, accessible internals with hatches etc. This makes the satellite more expensive and heavier. The difference was several hundred kg for Hubble.
No current spacecraft contain a manipulator arm or an airlock, so you can't use them. You could bolt the hardware onto an existing spacecraft, but only in an expendable fashion (attached to the service module).
The Space Shuttle had a large cargo bay that contained these items, so it could land with its repair/recovery tools intact.
You'd have to design a new spacecraft that includes these in a way that allows recovery of these expensive parts.
Lowering launch cost changes the economics, and makes it more feasible to design and launch a heavy capsule with spare parts, an airlock and a manipulator arm.
At the moment, satellite and launch are in the same cost bracket (\$100-300M each). When launch services become cheaper (say, \$10M) it would become cost-effective to service a satellite rather than replace it.