It's also not clear to me why self-knowledge isn't just a special case of person-knowledge, or even agent-knowledge. I know way more about myself (maybe) than anyone else, but it would seem to be the same sort of interpretation.

I agree that a separate category for self-centered memories doesn't make much sense, especially in the field of educational psychology, where theorists of memory (especially Baddeley) hold that memory (whether working or long-term) begins with perception. Perception (at least as they imagine it) is a rather self-centered process ANYWAY, so it seems that (again, in these theorists' models) memory is, by definition, self-centered.