At first glance, it seems that the illustration in the original question shows a much larger sideslip angle than could be maintained in the real world for the illustrated rudder deflection. I.e. the illustrated rudder deflection appears to be much too small for the illustrated sideslip angle, especially considering the ample size of the vertical fin in relation to the rudder. As illustrated, the rudder is almost streamlined to the undisturbed relative wind-- it would seem that it can't work that way!
Granted, the fin-fighting-rudder model outlined below is a little simplistic and if we look at the fin-rudder combination as a highly cambered airfoil flying at a mild negative angle- of-attack, perhaps we can see how it can still generate a net sideforce opposite to the direction of rudder deflection, even if the rudder is nearly aligned with (streamlined to) the direction of the undisturbed relative wind. (This appears to be one of the key points made by another answer.)
In that case the appropriate model would be not that the fin is "fighting" the yaw torque generated by the rudder, but rather that the fin/rudder combination is acting as a unit to "fight" the opposing sideforce and yaw torque generated by the rest of the aircraft, even though the (highly cambered) fin-rudder combination is actually flying at a slightly negative angle-of-attack in relation to its chord line. This idea is really just a tweaking, not a fundamental revision, of the other ideas expressed below.
For the rest of this answer we'll follow the more simplistic "fin fighting rudder" model:
But in a forced slip, isn't the fin's airfoil now in a negative
angle-of-attack (from the perspective of the earlier force). How is
the force maintained?
Yes, the plane will adopt a sideslip angle such that the nose-left yaw torque arising from the rightwards sideforce from the deflected rudder is exactly equal and opposite to the nose-right yaw torque arising from the leftwards sideforce generated by the vertical fin and other rear parts of the aircraft (minus the opposing yaw torque contributed from the forward parts of the aircraft.)
For simplicity, you can imagine that the fin and the rudder are "fighting" each other. If the fin is very large and the rudder is very small, only a small sideslip angle will be possible. If the fin is very small and the fuselage is very slender but there is a large rudder that can deflect a long ways, then a very large sideslip angle will be possible.
It's only by virtue of the fact that the rudder is BEHIND the fin-- or more accurately, behind the center of lateral area of the whole rest of the plane-- that we can be sure that left rudder deflection will in fact generate a NET leftwards sideforce in nearly all aircraft. The rudder is acting at a greater moment-arm than the rest of the plane, so if net yaw torque is to be zero, the rudder must be generating less sideforce than the rest of the plane, so the sideforce from the rest of the plane dominates over the sideforce from the rudder. That's why the slip-skid ball moves AWAY from the deflected rudder whenever net yaw torque is zero, i.e. yaw rotational acceleration is zero, as must be approximately true most of the time. (So we "step on the ball" to center it.)
If we are fighting an asymmetric effect like P-factor and we apply ONLY enough rudder to exactly center a yaw string, then we can see that since it is reacting to the rudder alone, and not the fin, the ball actually deflects slightly in the SAME direction as the rudder. That's why we leave the ball slightly deflected in the same direction as the rudder when we lose one engine in a twin--because it centers the yaw string. So we should apply a little less rudder than would be needed to fully center the ball. (Disclaimer-- nothing I say about twin/multiengine airplanes comes from personal experience.)
Imagine an exceptional case-- let's say we have a finless swept-wing flying wing aircraft, to which we add a rather short, very slender tail boom, and a large all-moving rudder with no vertical fin attached. Now we deflect the large rudder, which is located not so far behind the aircraft CG, 45 degrees to the left. Can we be so sure that the net sideforce will be to the left in this unusual case? The yaw string will certainly deflect to the left, but will the slip-skid ball really deflect to the right? Since yaw stability is not being provided by the sideforce from a rear fuselage and vertical tail, but rather from the differential drag and yaw torque generated by the swept wings in the sideways flow, it seems very likely that the rightwards sideforce from the rudder might dominate over all opposing sideforces, and the slip-skid ball might actually shift to the left. Meaning that if the wings are kept level, the flight path would actually curve to the right, not the left-- AWAY from the direction of the pilot's rudder input.
Now, think about what direction of bank would be needed to maintain a linear flight path when landing such an aircraft while holding a strong left rudder deflection to keep the landing gear aligned with the runway while landing with a strong right crosswind! The upwind wing would have to be raised, not lowered!
And also imagine what would happen if the pilot landed such an aircraft in a right crosswind using a wings-level, crabbed, non-slipping approach followed by a left rudder "kick out the crab" input just before touchdown. If the touchdown were delayed and the pilot kept holding/ adding more left rudder as needed to try to hold the nose on the runway heading, while applying roll inputs as needed to keep the wings level, the pilot would end up with the left rudder pedal all the way to the stop as the aircraft veered to the right off the runway heading, flying in a wings-level skidding circle in the clockwise direction, initially taking the plane beyond the UPWIND edge of the runway, and then continuing on through the rest of the circle-- rather weird!
Changing tack a little, there are aircraft where it is possible to maintain a pronounced sideslip with the pilot's feet completely off the rudder pedals. I've seen this in a Challenger ultralight, and in some sailplanes with small vertical fins. Here's this works: the glider or airplane is in right bank, with the yaw string streaming far toward the left, i.e. the slip-skid ball deflected far to the right. The sideways airflow hitting the fuselage generates a sideforce that prevents the flight path from curving toward the low wingtip. The sideways airflow also interacts with dihedral and other related effects to create a left roll torque tending to roll the plane towards wings-level, which the pilot is opposing with ample right aileron to maintain the bank angle. The deflected ailerons generate a strong "adverse yaw" torque, toward the left, which maintains the slip and keeps the yaw string deflected. In essence the ailerons are creating a strong "backwards rudder" effect in addition to their roll function. Typically the rudder has floated all the way over to the stop and remained there-- to the left in this case. Any attempt to actually turn right by increasing the right aileron deflection can simply end up increasing the slip angle (yaw string deflection angle) without changing the bank angle appreciably, possibly even forcing the flight path to curve toward the left. So, it's not always true that slip is maintained purely by the pilot's rudder inputs! Aircraft with flight characteristics like those described above are almost impossible to control with the ailerons and elevator alone-- constant rudder "coordination" inputs are essential, and simply taking one's feet off the rudder pedals and trying to control roll with the ailerons alone eventually tends to lead to the aircraft getting "stuck" in a sideslip due to adverse yaw.
