I only have **two minor wording changes** left. First, the abstract still calls Johansson's input a “rigorous Li table.” Since the body is now very careful about normalization, I'd make the abstract equally precise: > “Johansson's rigorous Keiper-coefficient table to \(10^5\), transferred to the Li normalization…” Second, Step 3 says: > “all other contour terms are unconditional.” I understand what you mean: they don't require RH at those heights. But because the formal contour proof consumes other trusted perimeter inputs, I'd make it narrower: > “all other contour terms are independent of the finite-height RH assumption.” or: > “no other contour term requires RH information above the residue cut.” That is harder for a referee to misinterpret.