I think you should treat separately usual translational forces and torques. Indeed the work done by a torque is equal to the difference of rotational kinetic energy. To keep things simple, assuming a constant torque, you get:
$$
\tau \, \varphi = \Delta K= \frac{1}{2}I\omega_f^2 - \frac{1}{2}I\omega_i^2
$$
where $\tau$ is the constant torque, $\varphi$ is the rotation angle, $I$ is Inertia momentum and $\omega_f$ , $\omega_i$ are final and initial angular velocities.
As far as your second point, if you carefully consider the orientations and the relative signs, the results will be the same, no matter which way you follow. For me, it's more intuitive to compute first the total net force and then compute the work.