Hi Dan,
My apologies for being too aspirational (if that’s the appropriate word), but I can’t read other people’s minds, so I couldn’t know what you like or don’t like. I just want this great project to succeed and improve.
However, I do know a few other things. One of them is that, in my experience, Shotcut is for most people the best and easiest way to learn professional-grade video editing without having to pay licensing fees or for expensive training. I also know from experience that the UI is generally well thought-out, and the learning curve is much lower than for, e.g., Avid, Premiere, Kdenlive or DaVinci. Now that you have added Bins and Subtitle Editing, it’s finally in the professional realm. In addition, it’s incredibly stable – I haven’t encountered a crash despite heavy CPU/GPU/memory use in the last two or three years.
From casual conversations I learned that, when I mentioned the virtues of Shotcut, one question was always: Can it be scripted for automation and other tasks? It turned out that advanced users wanted two features: 1) a scripting console for quick tasks, with the option to save the script (very important) and 2) a scripting interface for plug-ins. The latter is important, because ‘native’ plug-ins written in C, C++ or whatever language need to be compiled for each platform separately, whereas a Python or QML script works without compilation, as you know. In short: scripting console and easy plug-in options via scripts (Python seems to be the favourite).
As I’ve written in another post, this is not a challenge for you or any other current developer, so you don’t have to feel obliged to implement my suggestion. My hope is that someone picks this up and creates the functionalities I described, perhaps as a subject of an MA degree project in CS, or just because it’s an interesting challenge. There may also be commercial sponsors who need these functions and are willing to spend some money on it. Who knows!
I don’t know enough about the architecture of Shotcut, which seems to be a bit complicated (ffmpeg, MLT, Qt …), but I’d love to see some outsider sink his teeth into the whole thing and offer a workable solution for inclusion into Shotcut. It would then be up to you to decide whether this will become part of Shotcut or not. That’s the beauty of Open Source development. 
As I’ve written before: sowing seeds to be reaped later and perhaps unexpectedly. 