Cell culture based scenarios cannot explain the proline (P) amino acid addition (and subsequent glycosylation) in the (PRR-A-R) furin cleavage site of SARS-CoV-2; as these could only conceivably have evolved against an host animals immune system. (for a deep dive into the SARS-CoV-2 genome features, you can read this: https://medium.com/advances-in-biological-science/explained-the-hard-evidence-why-the-sars-cov-2-genome-was-not-engineered-and-unlikely-leaked-f3d6fbb10ff9)

Also, even if we assume serial passage in something like the humanized ACE2-receptor mice, RaTG13 has so many differences to SARS-CoV-2 that there would not have been enough time to create this divergence with just passaging. (Given that the lab is not that old, animal experiments have to be registered, especially such long-term ones, etc etc). No matter how you put it, lab leak theory does not work out without a lot of extra assumptions (about experiments, technologies, unrealistic scientific knowledge and competence, secret virus genomes and so forth). That's why it is still unlikely compared to zoonotic jump.