McDowell, knowledge and credit

I'm interested in getting clear about two features, specifically, of John McDowell's theory of knowledge. (1) How does his account answer any of the three value problems; (2) How does his account explain why knowers are credit-worthy?

I haven't yet come across literature that can explain the first of these questions, but after thinking about 'Knowledge and the Internal' I am somewhat confused as to how a satisfactory answer to (2) would go.

McDowell, in criticizing the 'hybrid' account, claims that what is epistemically significant between a knower and a non-knower (take the context of the New Evil Genius setting) should not fall outside the knower's standing in the space of reasons. And so, because the hybrid account posits that the 'favor from the world' is something external to the knower's standing in the SOR, McDowell rejects the view.

However, by preserving the Sellarsian idea of knowledge as a satisfactory standing in the space of reasons, McDowell abolishes the hybrid idea by Trojan Horsing in what was external in the hybrid account so that it falls within the knower's standing in the space of reasons. He does this by requiring that facts be reasons, and also be such that they shape the space of reasons in which agent finds herself.

What is troubling to me is this: McDowell, by making this move, thinks that facts are reflectively accessible. However, this just seems dubious to me when we think about an agent and his envatted counterpart forming the belief that p. What seems to be reflectively accessible to both is the content of some belief, which they take as a reason. For the lucky non-envatted agent, the content reflectively assessible to her is a fact. However, crucially, it is not reflectively accessible to the non-envatted agent that what she takes as a reason is a fact. And so, what is 'epistemically significant' between the knower and non-knower is something which, as in the case of the hybrid view, is not obiviously creditable to the agent. And so, I can't see how his account is going to explain why knowers are any more creditworthy than non-knowers.

The above is my 'naive' concern, and perhaps this will be assuaged after I get clearer on his view.

That aside, I'd be interested in either reasons for or against endorsing my concern and also any literature recommedations that would be useful to getting clear on what McDowell's view would say (on either of the initial questions)