So what did you do to support your assessment that the AGC computer code was not "capable?"
Experience. A few hours looking at the code, looking for algorithms that could "auto-pilot a spacecraft while maintaining attitude control". Inclusion of feedback loop for attitude control. Inclusion of "math to predict RCS response based on a changing Inertia Moment" (as fuel is used up, or AM vs. LM). Inclusion of Mascons, which cause a variance in steady-state orbital altitude by up to 7 nm, every 2 hours, and the impact this might have.
Lots of missing stuff. So then I stopped, because I knew this would be "hard-to-prove" and so futile.
43KHz speed for low-level instructions - doesn't help. 4KB RAM doesn't help. Archaic Computing tech doesn't help.
So without emulators - this is all just "unprovable suppositions" -- so a waste of my time. No fruit can be born from this without emulation of the system, including theoretical inputs from the dozens of other electronic sensors and devices attached.