C08.2 · Claims used

Registrar master of record (house-supplied identity): `cab1d95e8b1e6caad72a30cec35e7ff3bc5394cdc65354b2e17b3d65c90c298a`.

Read source: `D1258_REGISTRAR_RECORDS_S310.jsonl`, SHA-256 `2eab5ffff75a823c7a7c80ba050fca3a4116fffbaf6c4fd6de1943b9b705bfbc`. The full master bytes were not supplied, so its master digest is an attestation, not a recomputation. No live Registrar fetch was attempted.

**LIB2-053 — Frobenius compatibility fixes the finite trace ratio**

> On the reflected module 1 ⊕ 3 ⊕ 3̄, the registered Frobenius identity forces the trace weights a = b, uniquely fixing the readout r = 2 + b/(3a) = 7/3.

<!-- CORRECTION D11 S326 --> > **CORRECTION (2026-09-10, S326, D11).** The line below states `review_state=DRAFTED`. Every record this article cites has read **`IN_REVIEW`** since the S312 machine derivation (master `beb874fe3cdeb5aa` today), and under R65 no machine word is a review. **The point stands — review metadata is not clearance; the value it names does not.** Body unchanged (★H5).

Literal source labels: `math_status=NOT_PRINTED`; `attachment=NOT_PRINTED`; `scorecard_tier_word=null`; `house_tier=null`. `review_state=DRAFTED` is review metadata, not clearance.

**LIB2-107 — The signed metric-weighted residue cannot supply the positive seven count**

> For a subcollection of registered g-orthonormal Im Os channels, the signed residue equals the positive-minus-negative channel count and lies in [−4,+3], excluding 7 as that metric-weighted residue.

Literal source labels: `math_status=NOT_PRINTED`; `attachment=NOT_PRINTED`; `scorecard_tier_word=null`; `house_tier=null`. `review_state=DRAFTED` is review metadata, not clearance.

Both records are cited at the supplied master identity and their literal labels. The finite trace theorem and the signed-channel bound have different definitions and different hypotheses. Their separate proofs below do not turn either into a physical export theorem.