Consider how much should we constrain about counters in the private kernel

[sirasistant]

Currently, the private kernel trusts the counters issued by functions. We should consider wether:

• We should check that for any given call all its sideffects are within the call boundaries (start and end) (I think yes)
• We should check that for any given call all of its nested calls have assigned boundaries that don't overlap with each other
• We should check that for any given call all of the sideffects it generates don't fall into the boundaries of nested calls
• We should check that all counters are strictly increasing (no idea how to verify this in the kernel circuit)

[iAmMichaelConnor]

I think Leila summarised most of these checks in the draft spec too, if I'm not mistaken: https://yp-aztec.netlify.app/docs/circuits/private-kernel-initial#verifying-the-counters

[sirasistant]

After this PR: #6365 we are doing the first two bullet points of the list 🎉 . Doing the latter two shouldn't be necessary since all transient data that can be read is siloed, but it's something to keep in mind in the future.