To me, constraint forces did not make a lot of intuitive sense until I put an intuitive mental model under them.
By themselves, constraint forces are not about causation. They are about correlation. If you have such and such an acceleration, you must have this much force. If you have such and such a force, you must have this much acceleration. Shut up and calculate.
The mental model that I personally put under a constraint force (e.g an object on the end of a rope) is that it is all about a stable equilibrium. Instead of an ideal rope with infinite stiffness, give the rope just a tiny bit of springiness. Now, if the object on the rope strays a little bit beyond the rope limit, the tension force gets higher. The result is that the object accelerates inward. It moves back to the rope limit. Similarly, if the object on the rope strays a little bit inside the rope limit, the tension force gets lower. Left to its own device, the object coasts back out to the rope limit.
By itself, this just means that there is an equilibrium point. But not necessarily a stable one. From experience, we know that things do not bounce back as fast as they left. There is always a little friction. Or hysteresis. Or whatever you want to call it. The equilibrium is stable. For a hemp rope (as opposed to a rubber band), the equilibrium will be approached quite rapidly.
This intuitive model does not need to be very quantitative. Or even very correct. It just has to be good enough to keep the intuition happy while one shuts up and calculates.You might find that the concept of
Operational Amplifiers is relevant. It's the same kind of mechanism. You have an underlying stable negative feedback loop that does not have to be very precise in order to yield a result which precisely follows a mathematical model.