Count C3 share-encryption proofs by threshold share rows #1685

Audit reference: I41

What can go wrong

C3 validation counted proof rows using DKG encryption moduli, but honest Shamir shares use the threshold parameter rows. Honest bundles were therefore rejected with the wrong expected proof count.

When it happens

The mismatch appears when C3 share-encryption proofs are checked after DKG dispatch for a supported committee preset.

What changed

The dispatched DKG preset is resolved to its threshold counterpart and validation requires the canonical 2 + 2L C2/C3 proofs.

Why the change works

The count now comes from the same threshold share shape that generated the bundle, so honest proofs fit exactly and malformed multiplicities still fail.

Checks

The parameter-domain regression and full actor-level TRBFV flow pass.