The WIV simply doesn't hide samples for long enough for this to be a likely scenario.
By the way, it's estimated there's multiple hundred infections by novel coronavirus pathogens in China every year. So why would the much lower number of viral escapes be considered beyond it when we have the additional constrait of the sample not having been shared with anyone, whereas normally this is done?
4-6 weeks is lower than what we've seen abroad - in Italy between the likely patient zero in September 2019 and the first official case in late february, 5 months had gone.