Non-claims

M-D and M-E are not claimed. The M-C${}_{\mathrm{curv}}$ obstruction is graph-specific: it concerns the specified two-edge graph, the XVI-L prescription and the M.5 curvature targets; it is not a claim about gravity generally and not an obstruction of broad M-C. Enlarging the ledger (by non-gauge-invariant quadratic representatives, which may remain within broad Target I, or by gauge-fixing/ghost or further graph contributions) gives a different matching domain, not covered here.

The certification of Gate S-bv is a written check of XVI-J, eq:rational, on ${\mathbb{R}}^4\setminus \{0\}$ only: the predecessor $S$ is globally defined (XVI-J), and its position-space identification at the origin is not established here; no $i\epsilon$ prescription of the predecessors is modified, and the Euclidean motivation is not used as a proof.

The $546$ extra variational directions of the time-only ${\chi }_{\mathrm{I}}\equiv 1$ reading are spatial-only and non-Lorentz; they are not proposed as counterterms.

Proposition prop:to-actual is a conditional transfer, not a certification of actual fixed-data dimensions.

$B_{\mu }=0$ does not remove prescription dependence: lower-order residues still depend on $\varpi$.

The linear divergence identities of M.7 are free-field gauge identities, not interacting Ward identities; prescription covariance is not interacting-sector symmetry.

The double-plateau intersection numbers of the previous M.7 draft ($5/5/3$) and its statement that the curvature-square Euler tensor has nonzero divergence are withdrawn (Remark rem:M7-audit).

No $\epsilon \to 0$; no UV completion; no interacting BRST/Ward identity as proved; no all-orders renormalization.

The XI inverse ${G_{\mathrm{gr}}}^{\mathrm{R}}=-\mathcal{P}{E_0}^{\mathrm{R}}$ is not equated with the X-sheet ${{\Delta }_R}^{\mathrm{gr}}=-4\mathcal{P}{E_0}^{\mathrm{R}}$.

Preimage coefficients and collision dimensions are not physical Wilson coefficients.

Scale-dependent subtraction and scale-variation coefficients are not physical beta functions.

No release claim of XVI-L is altered.

Computational certificates support the stated identities; they do not replace the written proofs.