"Ensure stakeholders are aligned on what matters most and why"

But it is instructions for LLMs, right? A way to describe something that the humans know and the LLMs don't.

LLMs literally cannot be stakeholders, by definition.

For example, readability. Where are the developers out there saying "I am very willing to sacrifice a lot of readability to get even a small improvement on e.g. abstraction cleanliness", and sticking with it?

Or "performance can take a huge hit at the cost of being dead easy to read and reason about". Coming up with a list of abstractly good-sounding qualities is just prosocial signaling without knowing what you're willing to sacrifice. There should be a FUCKIT.md that enumerates these.