When suggested edits contain possible spam, the system adds a notice to the review task that says something like "hey, be careful -- the system identified this as possible spam" so that reviewers know to look more closely. Can we do something similar with comments and nearly-hidden text?
Recently we saw a suggested edit that included an inappropriate HTML comment. (It accused the author of trolling.) A reviewer who was looking at rendered text, not markdown source, missed it. And I know that I have sometimes missed the vandalism in edits where somebody wrapped a spam link around a single mark of punctuation (period or comma), and I didn't notice the teeny tiny link in the rendered text.
I think an edit containing any HTML comment or any link with super-short anchor text (a few characters or shorter) calls for additional care. It might be legitimate; those sorts of edits shouldn't be barred. But could we alert reviewers, like we do with possible spam? An automated system is already checking the edit; can it check for sneaky vandalism too?