The main solution to speed up stall recovery has been to install more powerful engines.
Stall is a total energy problem. If the aircraft is low, it has little potential energy left to top up its kinetic energy, and a stall is exactly when the kinetic energy drops below the value needed for staying in the air.
As you correctly note, the boundary layer needs some time to recover, and the further back on the wing you look, the more the boundary layer is shaped by what went on before. This helps to delay stall when the pitch rate is sufficiently high, but makes recovery take longer. Boundary layer suction would certainly help, but all cases of suction which I know of were either installed to keep the flow laminar for longer or to delay separation at high angle of attack. The suction volume needed to speed up stall recovery is a magnitude above that needed for delaying stall, so even when suction was installed, it would have been insufficient for shortening the stall recovery.
The idea of quickly adding kinetic energy is much more promising, and your proposal of a rocket engine will certainly help. If it points slightly upwards, it can compensate for the lift loss during stall and speed up the airplane so it returns quickly to the linear flight regime. The same can be achieved with sufficiently powerful engines, and since they are much more fuel efficient than rockets, they have been preferred over rockets to make fighter and aerobatic aircraft essentially unstallable if their thrust-to-weight ratio approaches one.
Next, a benign lift curve slope at stall can be engineered into the wing by using a big nose radius, a linear pressure recovery gradient and washout, so the outer wing still has attached flow when the inner wing stalls. The first two factors help to keep lift fairly constant well into the stall, so vertical acceleration is low. The third factor ensures proper aileron response and low roll moments due to flow separation, so the aircraft will not divert and stays controllable. The drag increase due to separation, however, will decelerate the aircraft and lead to a loss of lift if the stall is not ended quickly. The benign lift curve slope of the wing made the stall in case of AF447 so uneventful that the copilot never realized that he had pulled the aircraft into a stall.
However, most aircraft are flown such that they never enter into a stall. Instrumentation and warning devices will prevent even incompetent pilots from stalling the aircraft unless they intend to do so, eliminating the need for emergency rockets or oversized engines.