I am not sure whether I get your question but basically the spacer as general rule should be 20nt + 3nt (NGG) of the PAM if you are using Cas9. 5nt or 8 nt is to short and will not have enough speficity. 20nt is optimal cause it also help to be more specific nd reduce of high off-target activity. Do you have any specific reason to use such a short gRNA?
I want to do REMI (restriction enzyme mediated transformation) to achieve a multiple tagging of the genome with a small barcode, but I don't have a means to select...unless I express the restriction enzyme itself together with a marker! and since I routinely express Cas9 in the living system in question, I would prefer to deal with this nuclease than with another which probably would present some unpredictable idiosyncrasy.