Disappointing in its treatment of "fundamentalist" accounts of causation. Some non-problems are manufactured and then used as indictments of various theories of explanation. For instance, the "failure" of certain theories to account for negative causation is a problem Craver seeks to solve. Negative causation is when, say, (my example) a dam holds back water, thus "causing" the village not to flood. Thus if you ask "what caused that village to be there" part of the answer is the dam.
This is troubling, because everything that could but doesn't intervene to prevent something is then included in the explanation of why it happened. So it's not just, say, the construction workers who built the house that explain why it is there, an adequate explanation for how the house came into being is also explained by the aliens from Sector X who generously decided not to blow it up. To me this opens Pandora's box. Craver acknowledges the problem but implores us to be practical when judging explanatory relevance, but never offers any theoretical way to demarcate legitimate from frivolous sources of causation. This is an unsatisfying and untheoretical answer to a problem of theoretical making.
Craver also criticizes Wesley Salmon's notion of "physical necessity" as inadequate to deal with Hume's problem of induction with respect to causes, but replaces it with notions of casual interaction that he leaves completely unexplained (compared to the explaining he wants Salmon to do).
I think the discussion of levels is very good, and while I found the neutrality to Jaegwon Kim's doctrine of casual closure to be troubling, Craver does not reject it, and I think it can be read as largely complementary to Kim's views on mental causation.