Design proposal for a Keccak circuit that can handle multiple variable-length inputs.
Context
After redefining $$\mathsf{circuitID}(vk) = \mathsf{keccak}(vk)$$ we will be hashing
Fixed-length inputs: The Pedersen commitment $m$
Inputs with one variable length: The proof ID is $\mathsf{keccak}(cid, p_0, \ldots p_\ell)$
Inputs with two variable lengths: The circuit ID is $\mathsf{keccak}(vk.fixed, vk.s, vk.pedersen)$ and the last two inputs have variable lengths.
Summary