As a former programmer, I hesitate to even suggest the following, as it is likely far, far more complicated than the previous descriptions lead me to believe it might be.
But if NVDA is detecting that an error has occurred, and what would be sent to the NVDA developers (or add-on developers), and all of that content comes from the log file working backward from current logged error to the last one to extract what's needed, could this not be automated as a function (or, perhaps) add-on for the beta itself?
Based on what's been described, which I understand may not be comprehensive, grabbing that traceback should be relatively straightforward programatically and would result in far more data for developers than relying on manual reporting would.
[Now, let the explanations about why the above is, in actuality, impractical, begin. It just can't be this straightforward or someone would have done it.]
Brian - Windows 10 Pro, 64-Bit, Version 20H2, Build 19042
The depths of denial one can be pushed to by outside forces of disapproval can make you not even recognize yourself to yourself.
~ Brian Vogel