FINAL PRODUCER DISPOSITION -- 2026-10-18

Manuscript SHA256:
63f1a1a2400046d0083e8c654a5730a6239a713f09560fce83bcea06153b05b1
Analytic checkpoint SHA256:
c041807f92005436abdfae2c10ec583f17453903c4acbc7d7ec0d5d2763b9ef8

MANUSCRIPT PROOF
Fresh manuscript-specific independent assistant review found no material
mathematical or transcription defect, including mixed normalization, local
factors, all-n block lift, PVM signs, coherent conditioning, rational thresholds,
adaptive capture, stopped favorable witness, rejection caps and resource bounds.
The earlier complete analytic audit remains bound to the frozen GENERAL_PROOF.
The historical learner review uses an earlier threshold variant; the final
manuscript and combined source audit govern the rational thresholds.

ATTRIBUTION
Fresh actual Google and primary-version checks plus a separate manuscript
attribution audit found no material attribution blocker. The sampler/congruence
split, MUB identity, Clifford algebra and bootstrap are credited as known.
Hinsche's conceptual precedent and the pure-input versus mixed-input distinction
are explicit. Optional additions of the Eq27 change of variables and an additional
MUB-design reference were not needed for the stated self-contained derivation;
the reviewed manuscript was retained unchanged. The report records primary
publisher access limits for the 1989 MUB source. No exhaustive forward citation
graph or worldwide priority certificate is claimed.

NUMERICAL EVIDENCE
Fresh dense PVM replay passed. First full-suite replay completed eight cases
then hit a NumPy memory allocation failure in the final p=5,n=2 case; it was not
called a full pass. The unchanged scripts were rerun with OPENBLAS_NUM_THREADS=1
and OMP_NUM_THREADS=1. That complete suite passed: 178 states, 36 raw batch laws,
1010 exact Fraction scalar grid identities. Fresh results and full replay log
are included alongside the original evidence. These experiments are finite,
floating-point corroboration of analytic results, not a full learner simulation,
arbitrary-n computation, or proof premises. The wrapper runs in fresh scratch
and sets thread limits only in child processes.

COMPILATION AND PRESENTATION
Desktop project compilation succeeded. Tectonic compilation produced the final
nine-page PDF; the final log has no undefined reference, overfull or underfull
box warnings. All nine pages were rendered using Poppler and visually inspected
in three page groups. No clipping, overlap, missing glyph or unresolved citation
was observed. Character-bound checks supplement this inspection.

PRESERVATION AND AUTHORITY
The prior sealed qutrit folder's 40 files and ZIP are checked against the saved
baseline; see PRESERVATION_RECEIPT.json. This consolidated package is a successor
assessment draft in the same family. No qutrit source, ZIP or prior proof was
edited. Scott authorized preparation and delivery to Math Paper Publisher for
assessment and a report to him. No publication/upload/Archive release is granted.

These are producer-side assistant reviews and reproducibility checks. They do
not replace independent Publisher assessment or external human refereeing.
