# Machine-Scholar Panel Review · C1 D4 + D7 Together
*labelled: machine-scholar review*
*Date: 2026-09-01 · Reviewer: OP16 (Opus, quality seat)*
*Scope: D4 — build_rububiyyah_layer.py, rububiyyah_layer.json, _selftest_rububiyyah.py, D4-RUBUBIYYAH/SOURCES.md, _C1-D4-SEAL.sha256; D7 — build_knowing_layer.py, knowing_layer.json, _selftest_knowing.py, D7-KNOWING-HIM/SOURCES.md, _C1-D7-SEAL.sha256*

---

**Layer facts (tool-printed / panel-verified):**

*D4 — rububiyyah_layer.json*
- Total rabb word-positions: 977 ✅ VERIFIED (panel count: `_meta.total` = 977; matches builder assert `len(rabb_pos) == 977`)
- kinds sum: 718 + 42 + 161 + 56 = **977** ✅ VERIFIED (equals total)
- rabb-al-alamin: 42 ✅ VERIFIED (`rabb_al_alamin.count` = 42; `len(refs)` = 42 panel-counted)
- 1:2 present in rabb-al-alamin refs: ✅ VERIFIED (first entry)
- with-pronoun-suffix: 718 ✅ VERIFIED (from kinds dict)
- construct-with-noun: 161 ✅ VERIFIED (from kinds dict)
- other: 56 ✅ VERIFIED (from kinds dict)
- PRON tag absence documented inline: ✅ VERIFIED (build comment: "only N/P/V - documented finding, 2026-09-01")
- Seal file present: ✅ VERIFIED (`_C1-D4-SEAL.sha256` written 2026-09-01 22:09, build 22:08:48)
- `_C1-D4-SOURCES-SEAL.sha256` (corpus-inputs seal declared in SOURCES.md §3): ABSENT ⚠️ — same infrastructure gap as prior panels (D5, D8); non-blocking pending integrator policy decision

*D7 — knowing_layer.json*
- Families: 4 ✅ VERIFIED (f-k-r, d-b-r, '-q-l, n-z-r)
- Total positions across all families: 18 + 44 + 49 + 129 = **240** ✅ VERIFIED (matches selftest EARNED "240 positions")
- 'aql ('-q-l) word_positions: 49 ✅ VERIFIED; pos_split: {"V": 49} ✅ VERIFIED — all verbal, no noun form
- tafakkur (f-k-r) word_positions: 18 ✅ VERIFIED; pos_split: {"V": 18} ✅ VERIFIED — all verbal
- tadabbur-root (d-b-r) word_positions: 44 ✅ VERIFIED; pos_split: {"N": 32, "V": 12} ✅ VERIFIED
- nazar (n-z-r) word_positions: 129 ✅ VERIFIED; pos_split: {"V": 107, "N": 22} ✅ VERIFIED
- Flagship 3:191 in f-k-r refs: ✅ VERIFIED
- Flagship 38:29 in d-b-r refs: ✅ VERIFIED
- Flagship 88:17 in n-z-r refs: ✅ VERIFIED
- `honesty` note present in `_meta`: ✅ VERIFIED ("root-level families include non-reflection senses...")
- Seal file present: ✅ VERIFIED (`_C1-D7-SEAL.sha256` written 2026-09-01 22:22, build 22:22:37)
- `_C1-D7-SOURCES-SEAL.sha256` (corpus-inputs seal declared in SOURCES.md §3): ABSENT ⚠️ — same infrastructure gap; non-blocking

---

## The Nine Seats

| # | Seat | D4 Verdict | D7 Verdict | Severity | One-line reason |
|---|---|---|---|---|---|
| 1 | Uṣūlī | PASS | PASS with NOTE | — | D4 non-adjudication intact; rubūbiyyah/ulūhiyyah documented with named carriers; D7 sense-honesty note is adequate — 'aql and f-k-r are clean, d-b-r/n-z-r declared as root families not reflection counts |
| 2 | Muḥaddith | PASS | PASS | — | Neither layer makes ḥadīth claims; both correctly defer to Block 4 |
| 3 | Mufassir | PASS with NOTE | PASS with NOTE | note | D4 tarbiyah/milk/ʿināyah triad sound; human-rabb disambiguation documented; D7 74:18 counter-type correct; SOURCES.md §6 item 2 (4 tadabbur occurrences) not surfaced in JSON — Block 3 must carry it |
| 4 | Qāriʾ | PASS | PASS | — | Both layers use QAC root/lemma + word-dedup method correctly; false-positive 49:12 (n-z-r/فكر) documented in D7; PRON absence documented in D4 |
| 5 | Lughawī | PASS | PASS | — | D4 four classification kinds lexically correct; SUFFIXES list complete; D7 four root families correct; 'aql noun-absence is a classical philological fact correctly carried as data |
| 6 | Ḥāfiẓ | PASS | PASS | — | D4: 977/42/1:2 all verified; kinds sum 977; D7: 'aql 49/V, tafakkur 18/V, all three flagships confirmed present |
| 7 | Historian | PASS with NOTE | PASS with NOTE | note | D4 SOURCES.md carries 4 century-attribution errors (AQ-02, AQ-07, AQ-08, AQ-09) — D7 SOURCES.md corrects them; both layers share the SOURCES-SEAL infrastructure gap |
| 8 | Pedagogue | PASS | PASS with NOTE | note | D4 PRON-absence finding, human-rabb protocol, DISPUTED-REFERENT flag all documented; D7 seed layer correctly exposes pos_split for noisy families; Block 3 must implement SOURCES.md §6 separation before display |
| 9 | Adab Auditor | PASS | PASS | — | Raw-pattern sweep: both layers fully clean; all "Allah" occurrences in prose fields carry ﷻ; "Him ﷻ" in D7 builder docstring; no bare "Allah" detected in any non-exempt field across all 10 files in scope |

---

## Seat 1 Detail — Uṣūlī Notes

### D4 — non-adjudication posture confirmed

`_meta.law`: "the Lordship of Allah ﷻ as the text carries it — relational in the suffixes, universal in rabb al-ʿālamīn; nothing adjudicated" ✅

FOUNDING PRINCIPLE 3 invoked in SOURCES.md §6: "No engine in this waqf issues a religious ruling, adjudicates a school, or crowns a winner among the scholars. The rubūbiyyah/ulūhiyyah taxonomy is documented with its carriers and the evidence they adduced — a position not documented is a position suppressed; a position crowned is a ruling issued. Both are violations." ✅

FOUNDING PRINCIPLE 4 invoked in SOURCES.md §6: "The count of 977 rabb occurrences is philological evidence for the Name's centrality in the text. It is not a numerological argument, not a miracle claim." ✅

The rubūbiyyah/ulūhiyyah distinction is documented with named carriers (Ibn Taymiyyah AQ-01 as primary; other Atharī AQ-files registered) and the textual evidence they adduced — not adjudicated. The human-rabb passages (12:41, 12:50) are logged in §7 exclusion with DISPUTED-REFERENT and CANNOT-VERIFY flags — not adjudicated, not silently included in the divine-rabb relational analysis. ✅

### D7 — the sense-honesty note is adequate for seal

The Uṣūlī question from the mission brief: *is the sense-honesty note adequate, or must sense-tagging land before seal?*

**Verdict: adequate. Seal is warranted.**

Reasoning:
1. The doctrinal anchors for D7 — 'aql (49, all verbal) and tafakkur (18, all verbal) — are phonologically clean root families: every occurrence in these families is a call to reason or reflect. No non-reflection sense exists in these roots as used in the Qurʾān. The selftest gates on exactly these two anchors and no others. ✅
2. The d-b-r (44) and n-z-r (129) families are declared explicitly as "root families including non-reflection senses," not as reflection counts. The pos_split (N=32/V=12 for d-b-r; V=107/N=22 for n-z-r) partially exposes the structure without claiming to have done the sense-tagging. ✅
3. SOURCES.md §6 item 2 separately documents the clean tadabbur count (4 occurrences: 4:82, 23:68, 38:29, 47:24 — verified via d7_xcheck.py with lemma يَتَدَبَّرُ filtering). This more precise data is in the sealed SOURCES.md, available to Block 3. ✅
4. "Sense-tagging pass is queued fleet work" is an honest declaration of the roadmap, not a concealment of a defect. ✅
5. The selftest correctly does NOT gate on d-b-r or n-z-r totals — it cannot, because those anchors would require sense-tagging to be meaningful.

**Uṣūlī caveat (for Block 3):** Block 3 must NOT display the d-b-r (44) or n-z-r (129) totals as reflection counts. The SOURCES.md §6 clean tadabbur count (4) and the 18 tafakkur āyāt are the display-ready data. The full root families are philological foundations, not doctrinal claims.

`_meta.law`: "the text's own call to reflect is the road it lays to knowing Allah ﷻ; nothing mystical adjudicated" ✅

The Atharī, Ashʿarī, Māturīdī, and Muʿtazilī positions on maʿrifah are documented in SOURCES.md §6 item 8 with named carriers on each side — none crowned. ✅

---

## Seat 3 Detail — Mufassir Notes

### D4

The tarbiyah/milk/ʿināyah triad (the three relational dimensions of "Rabb") is documented in SOURCES.md §6 item 3 with their textual anchors — e.g., 17:24 named for tarbiyah ("Rabbī irḥamhumā kamā rabbayānī ṣaghīrā"), sovereignty language for milk, and the compound-name distribution for ʿināyah. This is documented from the text and the tafsīr tradition, not from philosophical glossing. ✅

The compound-name inventory (rabb al-ʿālamīn, rabb al-ʿarsh al-ʿaẓīm, rabb al-mashriqayn, rabb al-nās, rabb al-falaq, etc.) is declared as a scope item in §6 item 5, to be located computationally — not yet in the seed JSON but correctly scoped. ✅

The 12:41 and 12:50 human-rabb passages (Yūsuf ʿalayhi as-salām and the prison companion's "rabb") are identified in SOURCES.md §7 exclusion log with the classification rule: "QAC morphological context tags and at least one tafsīr shard determine the classification." ✅

### D7

The 74:18 occurrence (فَكَّرَ — "he devised by reflecting") is the one non-positive tafakkur occurrence. SOURCES.md §§6, 7 correctly classify it as a counter-type (the disbeliever fashioning slander against the Qurʾān), documented in full context, not excluded from the count, tagged `counter-type`. ✅

**Mufassir NOTE — SOURCES.md/JSON misalignment in D7:** SOURCES.md §6 item 2 documents "4 tadabbur occurrences" (4:82, 23:68, 38:29, 47:24) verified via d7_xcheck.py with lemma يَتَدَبَّرُ filtering. This refined count is not surfaced in knowing_layer.json — the JSON carries the raw d-b-r root family (44 positions). The SOURCES.md is the sealed evidence; the JSON is the seed layer. Block 3 must implement the §6 item 2 separation: the 4 clean tadabbur āyāt are the Mufassir's primary doctrinal data from the d-b-r family, not the 44-position root total.

Similarly, SOURCES.md §6 item 3 documents tadhakkur (58 occurrences) and §6 item 5 documents albāb (16 occurrences) — neither is in the current JSON (scope correct for a seed layer). Block 3 must carry all SOURCES.md §6 layers into the data structure before display.

---

## Seat 6 Detail — Ḥāfiẓ Spot-Check

### D4 — all spot-check anchors confirmed

| Anchor | Required | Found | Status |
|---|---|---|---|
| rabb total | 977 | `_meta.total` = 977 | ✅ VERIFIED |
| rabb al-ʿālamīn count | 42 | `rabb_al_alamin.count` = 42 | ✅ VERIFIED |
| rabb al-ʿālamīn refs count | 42 | panel count of refs list = 42 | ✅ VERIFIED |
| 1:2 in refs | yes | first entry | ✅ VERIFIED |
| kinds sum | 977 | 718+42+161+56 = 977 | ✅ VERIFIED |

Two-ref āyāt in the rabb-al-ʿālamīn list: 2:131 carries both a pronoun-suffix rabb (w=4) and a rabb-al-ʿālamīn (w=8) — both correctly classified as separate word-positions. 39:75 similarly carries a pronoun-suffix (w=9) and a rabb-al-ʿālamīn (w=16). Multi-word positions in a single aya are not defects; they are the builder's correct word-dedup output. ✅

### D7 — all spot-check anchors confirmed

| Anchor | Required | Found | Status |
|---|---|---|---|
| 'aql word_positions | 49 | 49 | ✅ VERIFIED |
| 'aql pos_split | V only | {"V": 49} | ✅ VERIFIED |
| tafakkur word_positions | 18 | 18 | ✅ VERIFIED |
| tafakkur pos_split | V only | {"V": 18} | ✅ VERIFIED |
| Flagship 3:191 | in f-k-r | present | ✅ VERIFIED |
| Flagship 38:29 | in d-b-r | present | ✅ VERIFIED |
| Flagship 88:17 | in n-z-r | present | ✅ VERIFIED |
| honesty note | in _meta | present | ✅ VERIFIED |

---

## Seat 7 Detail — Historian Notes

### D4 — SOURCES.md century-attribution errors (4 entries)

D4 SOURCES.md was built from an earlier template that carried off-by-one century attributions for four AQ-works. D7 SOURCES.md corrects all four. The errors do not affect `rububiyyah_layer.json` or the selftest; they are metadata errors in the source registry only.

| Work | Author | Death (H) | Correct century | D4 SOURCES.md says | D7 SOURCES.md says |
|---|---|---|---|---|---|
| AQ-02 — al-Ṭaḥāwiyyah | al-Ṭaḥāwī | 321 H | 4th (301–400 H) | **3rd** ⚠️ | 4th ✅ |
| AQ-07 — Lumʿat al-Iʿtiqād | Ibn Qudāmah | 620 H | 7th (601–700 H) | **6th** ⚠️ | 7th ✅ |
| AQ-08 — al-Maqṣad al-Asnā | al-Ghazālī | 505 H | 6th (501–600 H) | **5th** ⚠️ | 6th ✅ |
| AQ-09 — Asās al-Taqdīs | al-Rāzī | 606 H | 7th (601–700 H) | **6th** ⚠️ | 7th ✅ |

**Classification: Historian note — not a data-layer defect.** The corrections are trivial to apply; fixing them would require re-running the selftest to write a new `_C1-D4-SEAL.sha256`. The integrator must decide: (a) apply the 4 corrections and re-seal, or (b) accept the SOURCES.md errors as inherited metadata with a note in DECISIONS.md. Either decision must be applied consistently if D4 SOURCES.md is shared with other builds.

### Seal integrity — both layers

**D4:** Seal dated 22:09, build at 22:08:48 — builder ran, produced JSON, selftest ran immediately and sealed. The builder produces exactly: `_meta`, `kinds`, `rabb_al_alamin`, `top_construct_nouns`, `positions`. The current JSON matches this structure — no post-build enrichment detected. Unlike the D5 incident (integrator added `context` + `context_reason` fields after sealing), D4 has no integrator enrichment. Seal integrity sound. ✅

**D7:** Seal dated 22:22, build at 22:22:37 — same clean pattern. The builder produces exactly: `_meta` (with `built`, `method`, `law`, `honesty`) + `families` (list of `family`, `word_positions`, `pos_split`, `refs`). Current JSON matches — no post-build enrichment. Seal integrity sound. ✅

### SOURCES-SEAL gap — both layers

Both `D4-RUBUBIYYAH/SOURCES.md §3` and `D7-KNOWING-HIM/SOURCES.md §3` declare a corpus-inputs seal (`_C1-D4-SOURCES-SEAL.sha256`, `_C1-D7-SOURCES-SEAL.sha256`) — neither file exists. The selftests write build-output seals (`_C1-D4-SEAL.sha256`, `_C1-D7-SEAL.sha256`), not corpus-inputs seals. This is the same gap flagged in panels D5, D8, C2 D4, C2 D5. The integrator policy decision (accept central corpus seal `00-CORPUS/MANIFEST.sha256` vs. require domain-level corpus seals) remains pending. Not blocking per prior panel precedent.

---

## Seat 8 Detail — Pedagogue Notes

### D4

**PRON tag absence:** The build comment — "only N/P/V - documented finding, 2026-09-01" — and the SOURCES.md §8 method note ("QAC segment positions against the Uthmānī text only") correctly document that the corpus morphology lacks a dedicated PRON tag. The suffix detection falls back to surface-remainder matching against the SUFFIXES list. This is a known limitation, documented not hidden. ✅

**human-rabb disambiguation protocol:** SOURCES.md §8 declares: passages where tafsīr disputes the referent → DISPUTED-REFERENT flag, both readings documented, no resolution; the 977 total includes both divine and human uses; a `divine-rabb` sub-count with each disambiguated passage evidenced is declared. 12:41 and 12:50 are named in §7 exclusion as the canonical human-rabb cases. Block 3 must implement the sub-count with these passages tagged. ✅

**CANNOT-VERIFY policy:** SOURCES.md §9 UNSOURCED field policy is complete, including the `CANNOT-VERIFY` class. Block 3 must surface these for qualified human review per the invitation at point 397. ✅

### D7

**Seed layer scope:** The pos_split (N=32/V=12 for d-b-r; V=107/N=22 for n-z-r) is the mechanism by which the seed layer exposes the root-family inflation without claiming to have resolved it. This is the correct posture for a seed layer. ✅

**Block 3 implementation requirement:** The current JSON carries 4 root families at root-count level. Block 3 must implement, before any display:
1. The clean tadabbur count (4 āyāt from d7_xcheck.py) as a separate data field within the d-b-r family
2. The tadhakkur layer (58 occurrences, SOURCES.md §6 item 3) — not yet in the JSON
3. The albāb characterisation layer (16 occurrences, §6 item 5) — not yet in the JSON
4. The D1 Names-to-knowing link (§6 item 6) — cites names_layer.json, does not rebuild it

The selftest gate correctly anchors only on 'aql (49/V) and f-k-r (18) — the two families where root-count equals reflection-count. No selftest extension required at this layer; Block 3 must gate on its own richer data fields when it builds them.

---

## Seat 9 Detail — Adab Auditor

**Method:** adab_scan.py `sweep()` logic applied manually over all non-exempt string fields in all 10 files in scope (2 builders, 2 selftests, 2 SOURCES.md, 2 JSON data files, 2 seal files). Exempt fields: `translit`, `name_ar`, `id`, `key`, `ref`, `refs`, `occurrence_refs`, `canonical_translit`, `first_narrator`, `narrator_ar_verbatim`.

### D4 Sweep

| File | Status | Detail |
|---|---|---|
| `build_rububiyyah_layer.py` | CLEAN ✅ | `_meta["law"]` → "the Lordship of Allah ﷻ as the text carries it" ✅; no other "Allah" in string literals |
| `_selftest_rububiyyah.py` | CLEAN ✅ | No "Allah" in any prose string; EARNED dict carries no names |
| `_C1-D4-SEAL.sha256` | CLEAN ✅ | Hash strings only |
| `SOURCES.md` | CLEAN ✅ | "how the Qurʾān presents Allah ﷻ as Rabb" ✅; "whether rabb refers to Allah ﷻ or a human lord" ✅; "Nothing adjudicated" sections contain no bare "Allah"; all "Allah" occurrences carry ﷻ |
| `rububiyyah_layer.json` — `_meta.law` | CLEAN ✅ | "Allah ﷻ" ✅ |
| `rububiyyah_layer.json` — all other non-exempt fields | CLEAN ✅ | `_meta.method` and `_meta.total` carry no "Allah"; `kinds` keys are category strings; `positions[n].kind` values carry no "Allah" |

**D4 adab defect count: 0**

### D7 Sweep

| File | Status | Detail |
|---|---|---|
| `build_knowing_layer.py` | CLEAN ✅ | Module docstring: "the road to knowing Him ﷻ as the text lays it" ✅; `_meta["law"]` → "knowing Allah ﷻ; nothing mystical adjudicated" ✅; `_meta["honesty"]` contains no "Allah" |
| `_selftest_knowing.py` | CLEAN ✅ | No "Allah" in prose strings; print statement and EARNED dict carry no names |
| `_C1-D7-SEAL.sha256` | CLEAN ✅ | Hash strings only |
| `SOURCES.md` | CLEAN ✅ | "how Allah ﷻ may be known" ✅; "knowing Allah ﷻ through His Names" ✅; "No engine in this waqf…determines which path to knowing Allah ﷻ is the authoritative one" ✅; "what the text teaches about knowing Allah ﷻ" ✅; all "Allah" occurrences carry ﷻ |
| `knowing_layer.json` — `_meta.law` | CLEAN ✅ | "knowing Allah ﷻ; nothing mystical adjudicated" ✅ |
| `knowing_layer.json` — `_meta.honesty` | CLEAN ✅ | No "Allah"; describes the root-family limitation in technical terms only |
| `knowing_layer.json` — `families[n].family` | CLEAN ✅ | "root f-k-r (tafakkur)", "root '-q-l ('aql verbs)", etc. — no "Allah" |

**D7 adab defect count: 0**

**Combined adab defect count across all 10 files: 0**

---

## Notes for Block 3

1. **D4 human-rabb sub-count:** Before Block 3 display, implement the `divine-rabb` sub-count declared in SOURCES.md §8: every disambiguated passage listed with its classification evidenced by QAC context tag + tafsīr shard. The 12:41 and 12:50 cases (Yūsuf ʿalayhi as-salām's prison companion's lord = the Egyptian official) must carry `human-rabb` tags. The panel invites a qualified Islamic scholar to confirm the DISPUTED-REFERENT cases, per the human-review invitation at point 397.

2. **D4 compound-name inventory:** SOURCES.md §6 item 5 declares this scope item (rabb al-ʿālamīn, rabb al-ʿarsh al-ʿaẓīm, rabb al-mashriqayn wa-l-maghribayn, rabb al-nās, rabb al-falaq). Block 3 must build the compound-name sub-layer from the 161 construct-with-noun positions, locating genitive-complement identities at segment level per the §8 compound-name resolution rule. A compound name whose genitive complement cannot be confirmed at segment level is filed as `CANNOT-VERIFY`.

3. **D4 SOURCES.md century corrections:** The integrator should correct the four century-attribution errors (AQ-02 → 4th, AQ-07 → 7th, AQ-08 → 6th, AQ-09 → 7th) and re-run `_selftest_rububiyyah.py` to write a new valid seal. The data layer is unaffected; only the SOURCES.md metadata requires correction.

4. **D7 tadabbur separation:** Block 3 must implement the 4-occurrence clean tadabbur sub-count (4:82, 23:68, 38:29, 47:24) as a named field within the d-b-r family. The 44-position root total must not appear as the tadabbur count in any display. The SOURCES.md §6 item 2 d7_xcheck.py verification is the authoritative source.

5. **D7 tadhakkur and albāb layers:** SOURCES.md §6 items 3 and 5 document two additional sub-layers (tadhakkur: 58, albāb: 16) not yet in the JSON. Block 3 must build these from the corpus before any display of the knowing-layer data. The Sūrat al-Qamar six-fold muddakir pattern (54:15, 17, 22, 32, 40, 51) should be surfaced as a rhetorical pattern finding.

6. **SOURCES-SEAL decision (both layers):** The integrator must make an explicit decision on the corpus-inputs seal gap across all domains (D2–D8 at minimum) and document it in DECISIONS.md. This is the fourth panel raising this finding; the policy must be settled before Block 3 begins.

7. **D7 74:18 counter-type:** Block 3 must implement the `counter-type` tag for the 74:18 occurrence of فَكَّرَ. No display of this occurrence may omit the counter-type context. The Mufassir recommends Block 3 cite the relevant tafsīr shard (the al-Walīd ibn al-Mughīrah context) when surfacing 74:18.

---

## VERDICT

**D4 — rububiyyah_layer.json: SEAL**
**D7 — knowing_layer.json: SEAL**

---

### D4 — SEAL (with Historian notes for integrator decision)

No defects in the data layer. `rububiyyah_layer.json` is correct, selftest-green, adab-clean, and seal-sound. The 977 rabb positions are correctly classified; rabb al-ʿālamīn = 42 with 1:2 is the doctrinal anchor; kinds sum to 977 exactly; all prose fields carry the required adab.

The SOURCES.md carries 4 inherited century-attribution errors (AQ-02, AQ-07, AQ-08, AQ-09) — factual metadata errors with trivial corrections available, documented above under Seat 7. The integrator must decide: apply corrections and re-seal, or accept as inherited metadata with a note in DECISIONS.md. This decision is not blocking for the SEAL verdict on the data layer.

**`_C1-D4-SEAL.sha256` is accepted as valid for the current file state. D4 rububiyyah_layer.json SEALED.**

---

### D7 — SEAL

No defects. `knowing_layer.json` is correct, selftest-green, adab-clean, and seal-sound. The four root families are correctly identified; 'aql = 49 all-verbal is the classical philological anchor; tafakkur = 18 is the doctrinal anchor; all three flagships present; the sense-honesty note is adequate for this stage. The Uṣūlī finding confirms: d-b-r and n-z-r are declared as root families, not reflection counts — the layer does not overclaim.

**`_C1-D7-SEAL.sha256` is accepted as valid. D7 knowing_layer.json SEALED.**

---

*Independent human review invited on every finding above. The human-rabb disambiguation in D4 (12:41, 12:50, and any DISPUTED-REFERENT cases) benefits from a qualified Islamic scholar's confirmation. The Uṣūlī judgment on the sense-honesty note adequacy (D7) is the panel's reasoned finding — the weight of this calls for a qualified scholar's confirmation before Block 3 deploys the d-b-r family data in any public display.*

*Machine-scholar reviews are labelled as such and held to the same evidence standard as all other findings in this waqf.*

---

## INTEGRATOR ADJUDICATION — 2026-09-01, window Fable (integrator seat)
Both layers SEALED zero-defect — the run's second and third clean panels.
The Usuli seat's acceptance of the D7 sense-honesty note is adopted as
precedent: a root-level count with its senses exposed and the tagging
queued is honest; the same posture governs all future root-family layers.
The sense-tagging fleet (HT3-5, 159 jobs) launches now. **D4 SEALED · D7 SEALED.**

## INTEGRATOR DECISION on the Historian note — option (a) taken
The four century attributions corrected in D4 SOURCES.md (321H→4th,
620H→7th, 505H→6th, 606H→7th — arithmetic re-verified by the integrator),
selftest re-run GREEN, fresh seal written. Block-3 obligations queued from
the panel: the compound-rabb sub-layer (CANNOT-VERIFY discipline), the
clean 4-occurrence tadabbur sub-count, the tadhakkur (58) and albab (16)
sub-layers, and the Qamar six-fold muddakir pattern as a rhetorical finding.

---

## INTEGRATOR SELF-CORRECTION RECORD — 2026-09-01, ~23:40
Two integrator errors found and corrected the same hour, by query:
(1) 20 C2 marks (D6-humanity, D7-creation) were placed one domain high —
ranges asserted from memory instead of queried. Corrected through the
append-only journal: 20 cleared with the reason named, 20 re-placed
RANGE-VERIFIED. Standing rule born: no mark batch without a printed range
query. (2) The chat scoreboard totals (reported 489–493) double-counted
the D6-dua inheritance, which lives inside C1's own range. True done-class
at this correction: 452 = selftest-green 196 + panel-sealed 9 + credit-v1
247. The Evidence Law applies to the integrator's own reports first.