Oh yeah, I was actually wondering about that. Compared to where one can just keep cloning the successful survivors as in in-vitro, in-vivo would require far more accuracy. Didn't know those new developments were also underway. Very cool.

Ah yes, that's why indicator genes may be co-introduced, like GFP for instance, and sure functional testing for the gene products themselves can also be done.

It doesn't. The CRISPR-Cas9 complex itself only cleaves and removes a designated portion of DNA, and another matching/desired short linear DNA sequence is introduced into the vacancy left by the double strand break (DSB) induced by Cas9 cleavage and connected through ligase activity, either through homology-directed repair (HDR) or non-homologous end joining (NHEJ). The latter process may induce new insertions and deletions (indels) around the joining sequence due to sticky ends being required to form and thus cause inaccuracies, so if it can be done HDR would be preferred for introducing new genes. Typically though the processes would all be separate, the tools would all be packed in one synthesized plasmid. An internal ribosome entry site (IRES) sequence may be included in the plasmid for translation and expression of distinct gene segments after transcription, if more than one gene is to be introduced, or even just during the initial phase where different components may need to be translated at the same time. In in-vitro or in vivo studies sometimes an indicator gene like the green fluorescent protein (GFP) gene can be co-introduced to check whether the edit was successful. Overall as you may guess it can be trickier to introduce new gene elements compared to just cutting and inactivating genes.