The Bound Is Sharp: The Dilution Cannot Be Priced From the Record, and the First Ceilings Cannot Be Tightened From It Either

Assault on the Vacuum reported six targets attacked in one campaign. One result was derived, and it was negative: the marking strength κ_M is not identifiable from the record the program owns, proved by a rescaling symmetry, and the whole marking-side route to the cosmological constant Λ closed. At adopted grade the assault also bought the first finite bounds in the program's history, Λ ≤ 1, Σ_grain ≤ 121/160000 ħ, and c_Δ ≤ 0.0015125 — ceilings, not values, with the residual risk concentrated in one factor: the transport's dilution d = (ℓ_T/R_out)^2, whose form was derived and every candidate value of which died. This paper returns to that factor and derives the assault's second negative, at the same theorem grade as the first: d is not identifiable from the owned record, and the adopted ceilings are sharp against that record — no computation from owned quantities can tighten them. For every value of d in (0,1] there is a neck — an outer radius R_out = ℓ_T/√d — carrying exactly the record the program owns: the same profile shape, the same owned functionals, the same home-grain data. The record contains the torsion profile's shape and no distinguished point on it, which the assault's own extraction had already found; what is proved here is the consequence the assault did not draw: since every d is realised by an admissible neck with the identical record, the supremum of Λ = d·f_cell·(κ_M/ħ) over record-compatible necks is exactly the adopted ceiling, approached as R_out → ℓ_T from above and never exceeded. Tightening the bound is therefore provably equivalent to pricing the grip fraction at the rim, α_rim — a new input, not a computation waiting to be run. As supporting work, the assault's census for the sixth target (66 constructible speeds, no admissible rung) is extended to an expression search over the owned geometric constants: 207 constructible values, zero integer rung hits at a relative tolerance of 10^-9, the closest approach a relative miss of 3.7×10^-5 at 34.9987. The headline does not move: the vacuum catastrophe is NOT SOLVED / PARTIAL, and the producing-count is 0.