The primary aspect of the task--the task for which overt directions are given prior to commencing--appears to involve attention to visual detail, sustained attention, and some level of decision speed. So aspects of attention and processing speed are likely being assessed. If the target involves spelling, then it would also be assessing morphological knowledge. If the target is something else, then that other skill would obviously affect the scores as well. The single item asking questions about the content of the text may or may not even be scored. It is not unusual for those kind of items (especially when there is only one per passage) to be used more qualitatively, as a gross indication of whether the person was reading naturally, or focused only on the mechanical skill.
Implications: I don't think this is really a measure of photographic memory, as the example you give involves recognizing and transforming patterns that have previously been stored into long-term memory, not new memory encoding.
I am also very good at spellchecking, but I don't necessarily hold anything that I spellcheck in my memory any longer than it takes to mark the correction. If I am revising or editing a written product, I usually do multiple readings: one for grammar and mechanics (spelling, punctuation, capitalization, syntax), and one for meaning/language/style. Not always in that order. I can read for both at the same time, too, if I choose, but don't always, and the pace will be a little slower.
