So, I see a couple of things with this circuit that make me not want to have to do with it. I don't even know if I want to comment because I may give you wrong advice so take all of this with a grain of salt.
I don't know enough to tell if this circuit would work as it is, but I do know that with a supply voltage of 15V VR4 would not be doing anything
Here I believe that the diode is there for protection. The "Accel" signal comes directly from CRT so it may float due to strain charge to levels which would fry the opamp. ... at leat I think... the opamp input should never go lower than V- minus 0.5 but there it seems it may go as low as -3 V... so
perhaps they did just see no harm in leaving the Zener there and didn't want to take the time / effort needed to remove it.
I think this circuit works in a broader context. The feedback loop of the opamp involves some elements of the CRT itself and additionally it has some components (R18, C3) which are factory picked, so they must be selected depending on the CRT and / or other components.
It seems to me the additional zener you don't have was used to float the circuit to a certain level so that the input signal coming from CRT "Accel" could be referenced to the level set by the "Current limit" pot and at the same time integrated.
The output of this integrator sums up in current at the junction between CR3 and CR4 with another current which depends on Time/Div setting making up the amount of current which is sunk by Q12 which ultimately modulates the output gate amplifier.
The goal of this would be to put a limit on the writing gun current. I think they would choose at factory components by measuring the voltage at teh "Accel" grid given a certain cathode current for the specific CRT and in this case they perhaps saw that the mod was necessary to keep the circuit working within tolerance.
As a bonus you get a signal which compensates the focus at various intensity levels (thus "auto focus").
I think messing up with this circuit poses the risk that this very same circuit tries to mitigate: burning the storage mesh with excessive current: If I understand correctly the storage mesh is made of non-conductive material so it could be burned by a beam which is too strong. As of which value the current should be limited to... I have no clue and must be some parameter on the datasheet of the CRT which of course is not easy to obtain I guess. Even though you have the data, measuring the relation beween Accel grid voltage and cathode current is tricky due to the high voltages involved (but not impossible).