The light vortex tutorials site has a suggested PI workflow process. But yes I think the best is to develope your own over time. Like Rick I started with a simple order of operations, something easy to note down and understand to follow that will take " my data" and produce " an image". So it should be repeatable with no fuss and the way I work and capture its also my simple way of just chucking some data into PI and do a quick process to see if I got what I was after before going on to more lengthy processing.
My simple steps have rarely changed as I've changed gear, knowledge and skills. But its been the basis for my more serious processing workflow hich is very fluid for me. When a new feature or drastic feature improvement occurs in PixInsight I find it can cut down the need to run certain steps in my workflow or add more to it. I like to try to understand what each step in my workflow is trying to achieve as sometimes they are too subtle to notice at the time but allow for a better result in a later step. So I like to run through other peoples tutorials and think "oh, they're using this feature here to reduce most fine grain noise with little to no alteration of signal... hmm...." and I break out steps and plug them into my workflow and see how they go, or maybe features are explained better so I can now make more sensible setting adjustments with understanding. It's an iterative process and my workflow is workable regardless of the target mostly. its also easy to isolate a step to rerun it several times with small value changes instead of once with an aggressive value. Its far from perfect but for me I'm happy with how its evolved and how I can use the same workflow instead of having lots of workflows for star scapes, nebulosity, dslr planetary, galaxies. so my workflow is working well for me from the point I have a stacked integration frame. (So I can take flats lights darks bias etc and align and stack them in any software I like regardless of my capture gear as long as I get a single 64bit integraion file to start processing with. I also have capture/preprocessing workflows that vary depending on my capture gear and formats. Generally if I'm using a scope and imaging cam to do planets or lunar/solar surface they have their own processing workflows, otherwise its typically widefield captured on a tripod and my PI workflow kicks in there. But I recently read some info on PI wavelets that can help me put those into my regular workflow and maybe get absolutely everything processed the same way. My goal: one workflow to rule all my data.
I've tried to avoid "if nebulosity do these steps" in my workflow, I've found with good star masks I can keep in steps aimed at enhancing/brightening nebulosity and galaxies and they dont noticably effect a plain starfield image as I'm enhancing just above the noise but still in the black. Its satisfying to get a workflow you are happy with. Some operations have a side effect of cleaning noise at certain scales and others sharpen detail at other scales and what order you do them in can eliminate the need for a noise reduction operation or reduce it to certain scales at subtle values. experiment with what works for you and gives you satisfying results.
I'll stop rambling now