In the experience of the people reading this thread that have gone through the route of implementing their own customized harness with Pi... is is something very time consuming that ends up being a project on itself, or do you actually get time to _do_ things with it?
I'm asking because I tend to spend way too much time working on my tools and not enough time working on product I want to build with the tools... and I'm afraid a custom harness could become another rabbit hole for me.
If it takes you more than 2 weeks to stand up a usable harness that can improve itself (with or without you driving the harness), then use something off the shelf.
Otherwise, go all in.
Source: Daily driving bespoke harness since August.
I’m doing this less since they implemented MCP in pi, which is good for my day job.
For side projects, I now have a bunch of half-assed pi extensions as side projects..
IMHO, you should look for prior art if you really need it to get something done, otherwise, enjoy tinkering!