On Knowing/Saying To

First a question. Does anyone know of any literature (in epistemology or philosophy of language) dealing with sentences like either of the ones below?

1) Tom knows to go to the store.
2) Sara said to go to the bank.

It seems to me that sentence 1 should be of interest to epistemologists because it seems to exhibit something akin to the phenomenon that gets labeled "factivity" in the case of "knows that" ascriptions. By which I mean, just as it only makes sense for someone to utter "Tom knows that the store is open" if they themselves are of the opinion that the store is open, it only makes sense for someone to utter 1 if they stand in the some relevant approval/recommendation relationship to Tom's going to the store.

2 is interesting in the philosophy of language, because it seems to be the relevant way to report an instruction in indirect discourse. For instance, if Sara said, "Go to the bank" (to Jeff), but Jeff didn't hear her, and asked me what she said, I might report her utterance by saying 2. Insofar as some philosophers of language invoke considerations about indirect discourse as evidence for propositions (and/or as evidence about the nature of such propositions), sentences like 2 seem to be just as relevant when we raise questions about the existence/natures of instructions (understood as objects of a similar kind to propositions).

It is an interesting dissimilarity between 1 and 2 that a) only Tom can be the agent of "go to the store", while b) Sara is going to be generally dispreferred as a possible agent of "go to the bank" and c) the agent of "go to the bank" in 2 seems highly context sensitive (that is, there are readings corresponding to "Sara said for us to go to the bank", "Sara said for him to go to the bank", "Sara said for you to go to the bank", etc.)

I don't have any big "a-ha" thoughts on any of this yet, which is partially why I am hoping someone has written something about them. Both constructions seem deserving of attention, though.