That seems to me to have been reasonable prep for the metal shell. I assume that by ‘soap’ you mean a ‘dish soap’ that is actually a detergent, like Dawn, and not a hand soap.
You’re already using a stronger pickling solution than this thread recommended:
One reason in the past we have recommended comparatively gentle ‘picklants’ has been to avoid damage or dissolution of solder joints on preassembled detail, as on brass locomotives. That probably wasn’t the case for your shell.
Sanding would remove any ‘skin’ on the casting and raise some tooth, and you could address any flash at the same time, but you might also lose detail. I suspect that good old ‘quad-ought’ steel wool (0000) might tend to burnish the surface as much as ‘key’ it for mechanical adhesion. Probably the “best” solution would be some kind of media blasting – there are ‘kits’ that use an airbrush-size compressor for small pieces that would work with at least some of the ‘right’ media for Zamak prep. That raises a directionless matte finish that ought to hold primer well.
I am a ‘fan’ of baking the primer in the oven. I suspect you could use the pecan trick to avoid potential danger from fumes: preheat the oven to 250 to 275 F, then turn it off and put the parts inside. By the time the oven cools to a safe removal temperature, the “curing” should be good without any element heating. But don’t do this if you have a gas oven with standing pilots!
Little toaster ovens may be a course less expensive (to both wallet and marriage) but they do not have the thermal mass needed for pecanery. That implies that you let the self-etching primer dry its full recommended interval before baking it – and that you set the oven up in a well-ventilated space like a garage, or on an rextension cord outside a door.