I have read about as much about CRT as the OP, and have not studied history or even Liberal Art (whatever that is, drawing naked people?) Feel free to call this rubbish, it may well be.
However, I do know a thing or 2 about models, and the most important thing to understand about models is that they are
all wrong, but some are useful.
CRT is a model, the
theory of compensating wage differentials is a model, the
labour theory of value is a model, as are newtonian and quantum mechanics and general relativity. We know they are all "wrong", in the sense that they are not a full and infallible methods of predicting what will happen in the future based on the present. They are all abstractions that attempt to explain the real world, which is actually composed of millions of laws, billions of individual circumstances and trillions on inter-personal interactions, in a way that can be grasped by our little monkey brains. What makes them useful is to what extent this abstraction allows us to come up with interventions that make the world better.
My criticism of the academic field of CRT is that they are too nationalist, in that they seem to fall into the common american mistake of forgetting that the rest of the world exists. One of the most important features of any model is its generalisability, in that how well is explains situations outside of the situation from with the data that is was built from was gathered. It seems to me that CRT actually is MORE applicable in other countries that have a more recent history of racially distinct laws. Myanmar is the one that immediately springs to mind, but I also think India is another good example, and how well it fits countries like the UK, France, Australia and Russia would be very interesting. If I was a CRT academic, I would fit the model to every country in the world and demonstrate that it explains a varying amount of the inequality that exists around the world, and how that relates to the countries history. It seems that would be uncontroversial. Once that was done, the question changes from "Is CRT a valid theory" to "how much of the observed inequality is explained by this model", which is a whole lot less emotive and can be discussed objectively.
From what I've read implicit bias is mostly bs. Iirc in the early days they were trying to show it had some relevance but better put together studies showed the earlier ones were fatally flawed.
I am not sure what you are referring to, but there is
very strong evidence from fMRI studies in this area.