n index skip scan will read the interior branch blocks - looking for changes in values. It would not read just the right and the left - it would read some of the interior branch blocks to find out where the value of the leading edge column "changes"
it would not be able to take advantage of a check constraint to read the right/left - as it doesn't do that sort of processing.
So, it'll read interior blocks, branch blocks (navigational blocks) to determine when the value of that column changes and when it does - it will go down and process that part of the index structure all of the way to the leaf blocks, then it goes back to the branch block and continues looking for a value change and if it finds it - it goes down that part of the index again.