No worries! And no need to be hard on yourself. :)
Sounds like you had some cool ideas, and I'd be interested to see the physics-based approach you had in mind!
I'd discourage you from working in your own engine. It can work, but, from experience, if you are working on a game in an engine you built, what you actually end up working on is the engine, and the game never progresses. I recommend sticking with Godot, or using another existing engine if you prefer. You can achieve all the same fine-tweaking in Godot, without needing to do ALL of the fine-tweaking. It's worth it to build something on what is and improve it rather than making something new from scratch.
None of us got everything we wanted to do done, so we're all aware of the calendar issue. :D Hope your life events were at least positive and constructive.
If you are looking at doing a very "oil surfing" physics platformer, you might get more mileage out of looking at Tribes than Celeste. If I can think of something more appropriate or come across dev diaries that look useful, I'll let you know. :)