That is correct: An edit would automatically dispute the "very low quality" flag raised for a post, even if the edit doesn't make the post more understandable. The assumption is probably that, if somebody was able to edit the post, the post was not that "unlikely to be salvageable ."
Clearly, the author of a post has clear in her/his mind what s/he meant to ask, and at her/his eyes the post is perfectly understandable. An edit done from the author of the post should not be taken as to mean the post was salvageable.
To avoid these situations, there could be two solutions:
- Not considering a minor edit, when automatically disputing a "very low quality" flag, whoever edited the post
- Not considering the edits made from the author of the post, when disputing those flags
Between the two options, I would consider the first one. Suppose the author of the post completely re-writes the post because s/he noticed her/his English was completely obscure; the post is completely changed, and probably the flag is not anymore valid.
Vice versa, a user who edited a link because was not correct should not invalidate the flag.
If it is not desirable to do so all the times, it could be considered the score of the post, when automatically dismissing that flag. The "very low quality" flag is already raisable for answers with a score of 0, or negative; maybe the flag should not be automatically disputed if the score of the post is negative (or at least -2). I would consider a score lower than -2, as the user who flagged the post could first down-vote it; in the case the post's score was 0, that would mean to avoid an edit would dispute the flag.
It could also be the "very low quality" is disputed after the post is edited because the flag is generally misused. As far as I understood, the "very low quality" is not a substitute for the "not an answer" flag, or "not a real question" flag/closing reason; it rather means "this is not an answer/this is not a real question, and the user is not qualitatively contributing to the site."
I don't have any data about the number of times the "very low quality" flag has been manually disputed from a user, or reject from a moderator, to say that is a reason for which "very low quality" flags are automatically disputed after an edit.