Some of what you said on Choice Voting (the tactical voting in non-meek variations) is already noted on the page. It's also dependent on which variation of STV you use, which is why it's not fair to say it's a problem with Choice Voting, which is more or less advocacy for STV in general. Using Meek's Method it would be extremely difficult to tactically vote in an election and the difficulties far outweigh the risk of failure (it is actually NP-hard in this case to exploit the non-monotonicity of stv).
"if candidate A beats candidate B in an election, and additional votes are cast which rank canddiate A above B, then it is possible that the result is that candiddate A loses."
I'm not sure what you mean by this. If you mean that candidate A is still in the election in round N and candidate B is eliminated in round N with the ballot's associated with candidate B then distributed to candidates, it is then possible that in this same round candidate A would fall below the threshold and lose. This does not mean this is a failure of the voting system. There's nothing wrong with this behavior and it should be fully expected. From a very simplified point of view you could think of this as "candidate A is beating candidate B and then this causes candidate A to lose," but this would be an incorrect metaphor.
"For example, a candidate may be harmed if more people rank him higher (while all else is maintained equal), which shouldn't be the case in a perfect voting system."
Technically this may be true in some cases because STV is a round-by-round voting system but it is not exploitable, which is what is important. A statement like "a candidate may be harmed if more people rank him higher" makes it seem like this is actually going to happen and I believe is confusing. This requires more detail, I think. —PhilipNeustrom