Brandon Watson at Siris has a short post where he points out for apparently the nth+1 time that Plato's account of knowledge is not one that describes it as "Justified True Belief." He points out that at the locus classicus, the Theatetus 201c-210b, Socrates does indeed consider a model of "knowledge" that could conceivably be considered a JTB analogue, but that Socrates expressly rejects it. This is by most reckonings a mature middle-to-later dialogue, and it reaches a conclusion every bit as "skeptical" as the early and supposedly faithfully "Socratic" ones like the Lysis or the Euthyphro:
"So, Theatetus, neither perception, nor true belief, nor the addition of an "account' to true belief, can be knowledge."Sometimes I think that Plato had no express "doctrines" at all. Rather, he had a host of negative qualifications carving out the space for where a true doctrine would be, plus the hope of cultivating the realization (the famous spark leaping from soul to soul in the Seventh Letter) that would make the negative invert into a positive.