I thought that clutch friction accelerates one part and decelerates other one. Am I wrong?

I decided to use simplified clutch model for now. The slippage in simplified version is defined as follows:
if clutch pedal is almost released(the clutch engagement is above 0.9), than assume locked clutch.

if clutch pedal is half-way pressed, than assume slipping clutch. In this case the torque transferred is just multiplied by clutch engagement and the gearbox and engine angular velocities do not match.

if clutch pedal is almost fully pressed(engagement is below 0.05), than assume unlocked clutch and engine works seperate from rest of drivetrain.

And my physics system finally got working.