@Ababakar - welcome to MobileRead.
Yes as you will have read repeated through this thread there is no magic bullet for grabbing the ISBNs, and there will always be the odd situation where it either cannot find it or gets the wrong one if there are multiple. However these are the exception rather than the norm.
As for your cbr/cbz files - this plugin does not look for the word "ISBN" - the very first implementation did look for preceding words, however due to so many variations (and to cater for bad quality OCR scan errors) it now just looks for a sequence of numbers that start with the right prefix for ISBNs and validate as an ISBN. If it cannot find such in your comic books my guess is that they are images rather than text, which the plugin cannot scan. You can see from looking at the log as to what it text numbers it did attempt to match on, and you can always do a conversion to ePub to verify for yourself what "text" the plugin found available to scan (since for all but PDFs that is exactly what the plugin is doing in the background - silently doing a conversion to ePub and then scanning the html pages for text). If your comic shows up as an EPUB containing image files where the ISBN is then that proves the plugin will be unable to extract it.