With respect, I think you don't understand one of the design goals of Snap!. In the old days, in Scratch, people would create mods -- modified versions of Scratch itself -- and they would compete in terms of the number of blocks they added. (Primitive blocks, I mean, of course, since back then you couldn't write new blocks in Scratch itself.)
When we announced BYOB 3.0, the predecessor of Snap!, to the world, we took the opposite approach. We were proud that we added only eight blocks (to be exact, seven blocks and a button) to Scratch, and that was enough to give us custom blocks, first class procedures, first class lists, and with those, the ability for BYOB's users to write, in BYOB itself, any control structure or data structure they wanted.
Alas, we have since grown well beyond eight blocks. But it's still our policy to provide the minimum number of blocks needed to allow users to build whatever additional tools they need. (Our idea of minimality has expanded to include fast implementation of important algorithms that would be slow written in Snap!, such as higher order functions.)
So, when it comes to metaprogramming with hat blocks, probably all we'll add is the ability to ringify a hat block, or a script headed with a hat block. Users should then be able to use the existing metaprogramming facilities to do whatever they need. We'll do some experimentation to make sure we haven't forgotten some missing feature that really stands in the way of users manipulating hat blocks, but we won't add blocks just because they're useful, if they can straightforwardly be written with existing tools.