session 2026-07-29: REVIEWED-84 placed + PENDING-87 (order attestation) + PENDING-86 amended; spec v2.9.0 landed in chamber-library

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01WuMjg3ipEVa3n8CoSzoyvc

REVIEWED.md

@@ -856,4 +856,18 @@ brief. No work is blocked meanwhile; `wake-digest.py --brief` remains the fallba
 - **Scope affirmed as filed:** V-SCAN untouched (44 canonicals); no reference-converter proposal for V-DPDF (separate §V byte-census question, per REVIEWED-67); Harrison's earlier pipeline steps unaffected; born-digital-EPUB Docling lane untouched.
 - **Process note:** the 9 bare-scans and the 2 boundary-case files (Arcades Project; Haydn/Mozart/Beethoven) should be docketed as their own tracked items rather than remain prose-only inside this item, per the PENDING-55 residual-item precedent.
 **If corrections satisfied:** land as a MINOR supersession of the §Tiering & Fence evidence-tier table (new V-DPDF row, independence stated as a requirement) plus `graduation-spec.yaml` declared data (discriminator, thresholds, extractor pair). Prior version frozen byte-identical; bounded-diff proof at landing. Tag commits REVIEWED-83.
-**Awaiting:** steward placement, then the Q3 demonstration before Harrison's graduation stamp releases.
+**Awaiting:** steward placement, then the Q3 demonstration before Harrison's graduation stamp releases.
+
+## REVIEWED-84 — PENDING-87 — Order attestation: REVIEWED-83 Q3 corrected, not merely reformulated (design-gate passed with required conditions)
+**Date:** 2026-07-29
+**Decision:** AUTHORIZED — proceed per the ruling. This ruling corrects REVIEWED-83's own Q3, on the jurist's error, not the executor's.
+**Notes:**
+- **Q1 — the Q3 precondition as written is unsatisfiable — CONFIRMED, and withdrawn, not merely reformulated.** The coverage-based guard is structurally blind to block-level reordering (Part II(3), demonstrated against the real `verify_body_conservation.classify`, real canonical, discriminating control intact). This was already established (Eichmann pilot §7, 2026-07-19) and already ruled (REVIEWED-74, 2026-07-24: "the reordering residual is scoped to the STANDING Q3 order-blindness stamp-block ... demonstrated, Eichmann §7") — independently confirmed against REVIEWED.md, not taken from this package. The jurist's 2026-07-28 Q3 demanded an outcome already known in this repo to be impossible; owned as the jurist's gap.
+- **Q2 — independence confirmed; `held` does not apply — AFFIRMED, with a required forward condition.** Four extractors, zero shared libraries, and divergence (not agreement) correctly used as the proof. Unresolved: only docling recovers correct order on the adversarial fixture; the other three are independent but uniformly wrong. Any future `order_attestation:` proposal must not pair docling against itself as the order comparison — it needs a second, differently-implemented order-capable method, or an explicit argued fallback to eyeball-after-gate on the order axis specifically. Not required for this design-gate; required before that mechanism is ratified.
+- **Q3 — constitutional — AFFIRMED**, consistent with the independence ruling of 2026-07-28: what a comparison can and cannot attest binds at the level of the trust architecture; the measure's parameters are declared data.
+- **Q4 — tier-general — AFFIRMED**, already settled by REVIEWED-74 ("stands for both tiers"), not newly decided here.
+- **Q5 — mechanism not yet ratified; eyeball-after-gate invoked as the interim ceiling.** Measured exposure is 1 of 84 born-digital PDFs, and that one file already scores clean (0.995–0.999). The sample (11 books, simulated corruption) does not clear the bar for ratifying a new mechanical instrument against that exposure. Land the REQUIREMENT text now (Part V(A), essentially as drafted); do not ratify the `order_attestation:` mechanism yet; discharge blocking condition (a)'s order branch for the interim via its own named alternative — an explicit, argued acceptance of eyeball-after-gate — applied now to the single flagged book. The position-sensitive measurement stands as evidenced future work if the steward wants it built regardless of current exposure; that is a pacing call, not a ratification requirement.
+- **V-DPDF disposition (resolves REVIEWED-83's practical blocker):** order-blindness is not a V-DPDF-specific gap needing a new precondition — it is the pre-existing, tier-general stamp-block V-TEXT already operates under (REVIEWED-74). V-DPDF inherits it identically: promotions may proceed to `converted`; the `verified` stamp withholds on the order axis exactly as it already does elsewhere, discharged for Harrison by the eyeball-after-gate acceptance above, not by a new mechanism.
+- **Process note (not ruled):** a keyword-search gap over PENDING/REVIEWED contributed to the Q3 error as much as executor self-containment did. Fold into PENDING-86's options rather than opening a new item.
+**If AUTHORIZED:** Land the MINOR supersession adding Part V(A)'s requirement text beside blocking condition (a); record the explicit eyeball-after-gate acceptance for the order axis pending a ratified mechanism; release Harrison's `verified` stamp once that eyeball pass is done; V-DPDF's REVIEWED-83 status otherwise unchanged (Q1, Q2, Q4, Q5 of that ruling stand). Tag commits REVIEWED-84.
+**Awaiting:** steward placement; steward eyeball pass on the one flagged book; steward pacing call on whether `order_attestation:` gets built now or banked.