# PANEL — C7 PILLAR COHERENCE REVIEW
**Reviewer:** OP50 (Opus, quality seat)
**Date:** 2026-09-02
**Mission:** Whole-pillar panel — does Pillar C7 (The Prophets & Messengers عليهم السلام) cohere across all nine domains?
**Scope:** D1-REGISTRY · D2-SECOND-TEN · D3-FINAL-FIVE · D4-PATIENCE · D5-DAWAH · D6-GRAPH · D7-IN-TAFSIR · D8-GEOGRAPHY-CHRONOLOGY · D9-ATLAS
**Points:** 5401–6300
**Verdict required:** PILLAR-COHERENT (with open-work register) or FIX-FIRST

---

## VERDICT

> ### 🔴 FIX-FIRST
> Four defect categories are present. Two are **gated adab defects** (bare prophet mentions without the required honorifics, including Muhammad without ﷺ, across D3, D6, and D8). One is a **data-integrity defect** (D6 graph presents homonym-contaminated raw counts for Hud عليه السلام and Salih عليه السلام as if they are clean prophetic-referent ayah counts, directly contradicting D1's declared HOMONYM-MIXED law). One is an **honest-documentation defect** (D2 is titled "second ten" but carries nine prophets; Dhul-Kifl عليه السلام is absent from all three registry domains with no explicit holding note). None of these are Block 3 open-work items — they are build-layer defects that must be resolved before any domain in this pillar can be sealed.

---

## §1 — LAYER FACTS (tool-printed)

All counts below are read from the sealed layer files on disk, not from memory.

### D1-REGISTRY (`first_ten_layer.json`)
Built 2026-09-02T00:10:06. Gates: 2026-09-02T16:02:43.

| Prophet عليه السلام | word_positions | unique_ayah_count | homonym_mixed | Label |
|---|---|---|---|---|
| Adam عليه السلام | 25 | 25 | false | ✅ VERIFIED |
| Idris عليه السلام | 2 | 2 | false | ✅ VERIFIED |
| Nuh عليه السلام | 43 | 43 | false | ✅ VERIFIED |
| Hud عليه السلام | 10 | 10 | **true** | ✅ VERIFIED — raw probe; ⚠️ HOMONYM-MIXED (hudan = guided/Jews in 2:111, 2:135, 2:140) |
| Salih عليه السلام | 74 | 73 | **true** | ✅ VERIFIED — raw probe; ⚠️ HOMONYM-MIXED (salih = righteous adjective in majority of refs) |
| Ibrahim عليه السلام | 69 | 63 | false | ✅ VERIFIED — 69 positions, 63 unique ayat (6 ayat carry name twice) |
| Lut عليه السلام | 27 | 27 | false | ✅ VERIFIED |
| Isma'il عليه السلام | 12 | 12 | false | ✅ VERIFIED |
| Ishaq عليه السلام | 17 | 16 | false | ✅ VERIFIED — 17 positions, 16 unique ayat |
| Ya'qub عليه السلام | 16 | 16 | false | ✅ VERIFIED |

**D1 total prophets: 10 ✅ VERIFIED — `len(rows)=10`**
**D1 seal: ⚠️ NOT SEALED — no prior independent panel; this is the first review**

### D2-SECOND-TEN (`second_ten_layer.json`)
Built 2026-09-02T14:55:31. Gates: 2026-09-02T16:13:00.

| Prophet عليه السلام | positions (word) | unique_ayah_count | Label |
|---|---|---|---|
| Yusuf عليه السلام | 27 | 26 | ✅ VERIFIED — 27 positions, 26 unique ayat |
| Ayyub عليه السلام | 4 | 4 | ✅ VERIFIED |
| Shu'ayb عليه السلام | 11 | 10 | ✅ VERIFIED — 11 positions, 10 unique ayat |
| Musa عليه السلام | 136 | 99 | ✅ VERIFIED — 136 positions, 99 unique ayat |
| Harun عليه السلام | 20 | 20 | ✅ VERIFIED |
| Dawud عليه السلام | 16 | 16 | ✅ VERIFIED |
| Sulayman عليه السلام | 17 | 16 | ✅ VERIFIED — 17 positions, 16 unique ayat |
| Ilyas عليه السلام | 3 | 3 | ✅ VERIFIED |
| al-Yasa' عليه السلام | 2 | 2 | ✅ VERIFIED |

**D2 total prophets: 9 ⚠️ — domain title says "second ten"; actual count is 9; Dhul-Kifl عليه السلام is absent (see §4 Finding 4)**
**D2 seal: ⚠️ NOT SEALED**

### D3-FINAL-FIVE (`final_five_layer.json`)
Built 2026-09-02T15:03:40. Gates: 2026-09-02T15:17:17.

| Prophet | positions | unique_ayah_count | Adab | Label |
|---|---|---|---|---|
| Yunus | 4 | 4 | ❌ NO عليه السلام | ✅ count VERIFIED — ⚠️ ADAB DEFECT |
| Zakariyya | 7 | 6 | ❌ NO عليه السلام | ✅ count VERIFIED — ⚠️ ADAB DEFECT |
| Yahya | 5 | 5 | ❌ NO عليه السلام | ✅ count VERIFIED — ⚠️ ADAB DEFECT |
| 'Isa | 25 | 25 | ❌ NO عليه السلام | ✅ count VERIFIED — ⚠️ ADAB DEFECT |
| Muhammad | 4 | 4 | ❌ NO ﷺ | ✅ count VERIFIED — ⚠️ ADAB DEFECT |

Note field also carries bare "Muhammad" without ﷺ.

**D3 total prophets: 5 ✅ VERIFIED — `len(prophets)=5`**
**D3 seal: ⚠️ NOT SEALED — adab defects present (see §3 Finding 1)**

### D4-PATIENCE (`patience_cooccurrence_layer.json`)
Built 2026-09-02T15:18:25. Method: co-occurrence only (sabr+balw+ftn roots ∩ prophet registries).

| Prophet | n | Co-occurrence ayat | Label |
|---|---|---|---|
| Adam عليه السلام | 2 | 20:120, 7:27 | ✅ VERIFIED — unique lemma, clean |
| Idris عليه السلام | 1 | 21:85 | ✅ VERIFIED — unique lemma |
| Salih عليه السلام | 3 | 18:82, 28:80, 7:168 | ⚠️ HOMONYM-CONTAMINATED — 18:82 = "abuhu salihan" (adjective); 28:80 = "amila salihan" (adjective); 7:168 = "al-salihin" (adjective). All 3 co-occurrences are adjective uses, not the prophet. Row is likely spurious (see §3 Finding 3) |
| Ibrahim عليه السلام | 1 | 2:124 | ✅ VERIFIED — unique lemma |
| Isma'il عليه السلام | 1 | 21:85 | ✅ VERIFIED — unique lemma |
| Yusuf عليه السلام | 1 | 12:90 | ✅ VERIFIED |
| Musa عليه السلام | 5 | 10:83, 14:5, 14:6, 20:40, 2:61 | ✅ VERIFIED — unique lemma |
| Harun عليه السلام | 1 | 20:90 | ✅ VERIFIED |
| Dawud عليه السلام | 2 | 38:17, 38:24 | ✅ VERIFIED |
| Sulayman عليه السلام | 2 | 2:102, 38:34 | ✅ VERIFIED |

**D4 rows: 10 ✅ VERIFIED — atlas confirms patience_rows=10**
**D4 seal: ⚠️ NOT SEALED — Salih row homonym-contaminated (see §3 Finding 3)**

### D5-DAWAH (`dawah_cooccurrence_layer.json`)
Built 2026-09-02T15:18:25. Method: co-occurrence only (da'w+balagh+nadhr roots ∩ prophet registries).

| Prophet | n | Suspect ayat | Label |
|---|---|---|---|
| Nuh عليه السلام | 5 | — | ✅ VERIFIED — unique lemma |
| Salih عليه السلام | 6 | 18:82, 35:37, 41:33, 46:15, 7:189 | ⚠️ HOMONYM-CONTAMINATED — 5/6 ayat carry adjective salih (righteous), not the prophet's name; only 11:62 ("ya Salihu qad kunta fina") is the prophet. Row is largely spurious (see §3 Finding 3) |
| Ibrahim عليه السلام | 2 | — | ✅ VERIFIED |
| Lut عليه السلام | 1 | — | ✅ VERIFIED |
| Isma'il عليه السلام | 1 | — | ✅ VERIFIED |
| Ishaq عليه السلام | 1 | — | ✅ VERIFIED |
| Musa عليه السلام | 5 | — | ✅ VERIFIED — unique lemma |
| Zakariyya عليه السلام | 1 | — | ✅ VERIFIED |
| Yahya عليه السلام | 1 | — | ✅ VERIFIED |
| 'Isa عليه السلام | 1 | — | ✅ VERIFIED |

**D5 rows: 10 ✅ VERIFIED — atlas confirms dawah_rows=10**
**D5 seal: ⚠️ NOT SEALED — Salih row homonym-contaminated (see §3 Finding 3)**

### D6-GRAPH (`prophet_graph_layer.json`)
Built 2026-09-02T15:15:33. Joins all three registry domains.

**Node count: 24 ✅ VERIFIED — `d['prophets']=24`**
**Total edges: 527 ✅ VERIFIED — `d['total_edges']=527`**

Edge check: sum of all node ayah_counts =  
D1 subset: 25+2+43+10+73+63+27+12+16+16 = **287**  
D2 subset: 26+4+10+99+20+16+16+3+2 = **196**  
D3 subset: 4+6+5+25+4 = **44**  
Total: 287+196+44 = **527 ✅ CALCULATED — matches total_edges**

| Adab state in D6 | Prophets affected |
|---|---|
| ✅ Honorific present | All D2 prophets (9): "Yusuf 'alayhi as-salam" etc. |
| ❌ Honorific ABSENT | All D1 prophets (10): "Adam", "Nuh", etc. bare |
| ❌ Honorific ABSENT | All D3 prophets (4): "Yunus", "Zakariyya", "Yahya", "'Isa" bare |
| ❌ ﷺ ABSENT | Muhammad: "Muhammad (see C4 D1 - the Final Messenger)" — no ﷺ |

⚠️ D6 also presents Hud ayah_count=10 and Salih ayah_count=73 without the homonym_mixed flag. These counts are raw probe figures, not referential-pass-confirmed prophetic occurrences (see §3 Finding 2).

**D6 seal: ⚠️ NOT SEALED — adab defects and homonym-contamination defect**

### D7-IN-TAFSIR (`prophets_tafsir_layer.json`)
Built 2026-09-02T15:18:25. Method: co-occurrence with consensus engine; cache_size=94.

Registered verses: 2 — `21:25` ("the one message of all messengers", required_elements=5) and `33:40` ("the seal of prophethood", required_elements=11).

The cache_size=94 indicates 94 consensus engine records exist; only 2 key cross-pillar verses are explicitly registered in this layer. This reflects the early-stage posture of D7: the consensus engine has produced the underlying data, but selection and registration for the domain layer is not yet complete. The layer is honest about its state (no overclaims). Routed to open-work register.

**D7 seal: ⚠️ NOT SEALED**

### D8-GEOGRAPHY-CHRONOLOGY (`chronology_layer.json`)
Built 2026-09-02T15:20:48. Source: the three sealed registries.

Canonical sequence: 10 (D1) + 9 (D2) + 5 (D3) = 24 prophets, same composition as D6.

D3 prophet entries in D8 carry bare names without honorifics: "Yunus", "Zakariyya", "Yahya", "'Isa", "Muhammad (see C4 D1 - the Final Messenger)" — the last without ﷺ. Same pattern as D3 itself — the adab defect propagated from D3 into D8.

Geography status: "AWAITED-CLASSICAL: place attributions (Makkah, Ur, Madyan, Misr, al-Quds) enter only with named classical sources" — ✅ HONEST.

**D8 seal: ⚠️ NOT SEALED — adab defects propagated from D3**

### D9-ATLAS (`atlas.json`)
Built 2026-09-02T15:23:57.

| Field | Value | Label |
|---|---|---|
| summary.first_ten | 10 | ✅ VERIFIED — matches D1 `len(rows)=10` |
| summary.second_ten | 9 | ✅ VERIFIED — matches D2 `len(prophets)=9`; ⚠️ labeled "ten" but has 9 (Dhul-Kifl gap) |
| summary.final_five | 5 | ✅ VERIFIED — matches D3 `len(prophets)=5` |
| summary.patience_rows | 10 | ✅ VERIFIED — matches D4 `len(rows)=10` |
| summary.dawah_rows | 10 | ✅ VERIFIED — matches D5 `len(rows)=10` |
| summary.graph.prophets | 24 | ✅ VERIFIED — matches D6 `d['prophets']=24` |
| summary.graph.edges | 527 | ✅ VERIFIED — matches D6 `d['total_edges']=527` |
| summary.tafsir_verses_cached | 2 | ✅ VERIFIED — matches D7 `len(verses)=2` |
| summary.chronology | "AWAITED-CLASSICAL" | ✅ VERIFIED — matches D8 `geography_status` |

**D9 seal: ⚠️ NOT SEALED — first panel review; defects in source domains propagate here**

---

## §2 — HAFIZ SEAT: KEY-COUNT RECONCILIATION

Counts that must agree wherever cited across the pillar. All confirmed by tool-read.

| Count | Mission anchor | D1 | D6 graph | D8 | Status |
|---|---|---|---|---|---|
| Ibrahim عليه السلام unique ayat | 63 | unique_ayah_count=63 ✅ | ayah_count=63 ✅ | in sequence ✅ | **CONSISTENT** |
| Nuh عليه السلام occurrences | 43 | word_positions=43, unique=43 ✅ | ayah_count=43 ✅ | in sequence ✅ | **CONSISTENT** |
| Adam عليه السلام occurrences | 25 | word_positions=25, unique=25 ✅ | ayah_count=25 ✅ | in sequence ✅ | **CONSISTENT** |
| Musa عليه السلام total positions | 136 | D2: positions=136 ✅ | — (ayah_count=99 = unique) ✅ | in sequence ✅ | **CONSISTENT** |
| Musa عليه السلام unique ayat | 99 | D2: 99 occurrence_refs ✅ | ayah_count=99 ✅ | — | **CONSISTENT** |
| 'Isa عليه السلام unique ayat | 25 | D3: 25 occurrence_refs ✅ | ayah_count=25 ✅ | in sequence ✅ | **CONSISTENT** |
| Zakariyya عليه السلام unique ayat | 6 | D3: 6 occurrence_refs (positions=7) ✅ | ayah_count=6 ✅ | in sequence ✅ | **CONSISTENT** |
| Total graph nodes | 24 | D1(10)+D2(9)+D3(5)=24 ✅ CALCULATED | prophets=24 ✅ | 24-prophet sequence ✅ | **CONSISTENT** |
| Total graph edges | 527 | D1(287)+D2(196)+D3(44)=527 ✅ CALCULATED | total_edges=527 ✅ | — | **CONSISTENT** |

**All nine counts are consistent across every domain that cites them.** No primary count contradicts another domain's assertion.

---

## §3 — DEFECT REGISTER (pre-seal)

### DEFECT 1 — ADAB: Missing honorifics in D3, D6, and D8

**Scope:** D3 `final_five_layer.json`, D6 `prophet_graph_layer.json` (D1 and D3 nodes), D8 `chronology_layer.json` (D3 entries)

**D3 violations (gloss field):**
| Prophet | Actual | Required |
|---|---|---|
| Yunus | `"gloss": "Yunus"` | `"Yunus عليه السلام"` |
| Zakariyya | `"gloss": "Zakariyya"` | `"Zakariyya عليه السلام"` |
| Yahya | `"gloss": "Yahya"` | `"Yahya عليه السلام"` |
| 'Isa | `"gloss": "'Isa"` | `"'Isa عليه السلام"` |
| Muhammad | `"gloss": "Muhammad (see C4...)"` | `"Muhammad ﷺ (see C4...)"` |
| D3 note | `"the Final Messenger Muhammad"` | must carry ﷺ |

**D6 violations (prophet node field):**
All 10 D1 prophet nodes bare ("Adam", "Nuh", "Ibrahim" etc.); all 4 D3 non-Muhammad prophet nodes bare ("Yunus", "Zakariyya", "Yahya", "'Isa"); Muhammad node bare (no ﷺ). D2 prophet nodes correctly carry "Yusuf 'alayhi as-salam" etc. — so the graph has inconsistent honorific practice across its 24 nodes.

**D8 violations:** D3 prophet entries (5) carry the same bare forms as D3.

**Fix required:** Apply عليه السلام to each prophet's gloss/node field in D3, D6 (D1 and D3 nodes), and D8. Apply ﷺ to all Muhammad references in these files.

**Severity: GATED DEFECT** — Adab Law is a standing panel check; a missing honorific is a gated defect that blocks the seal.

---

### DEFECT 2 — DATA INTEGRITY: D6 graph propagates raw homonym counts without flag

**Scope:** D6 `prophet_graph_layer.json` — Hud node (ayah_count=10) and Salih node (ayah_count=73)

**The conflict:** D1 declares both as HOMONYM-MIXED and states explicitly: *"HOMONYM-MIXED: raw probe = 10 word positions; lemma hud/هود also carries hudan (guidance/the guided people) and the tribal name; referential pass (two-shard) required before any per-prophet figure is published."* and the same for Salih. D1 records `homonym_mixed: true` in both rows.

**D6 violates this by:**
1. Assigning `ayah_count=10` to the Hud node — this figure is presented as the node's authoritative ayah count, but includes at minimum 2:111, 2:135, 2:140 which carry *hudan* = "the Jews / those guided", not the prophet.
2. Assigning `ayah_count=73` to the Salih node — the majority of the 73 occurrences are the common adjective *ṣāliḥ* (righteous), not the prophet's name (e.g. 2:62 "wa-'amila ṣāliḥan", 3:114 "min al-ṣāliḥīn", 18:82 "wa-kāna abūhu ṣāliḥan").
3. No `homonym_mixed` flag on either node — the graph reader has no way to know these counts are raw, not referential-pass-confirmed.

**Fix required:** Either (a) add a `"homonym_mixed": true` field and an explicit note on the Hud and Salih nodes declaring the counts are pre-referential-pass, or (b) remove both nodes from D6 until the referential pass has been run and the clean per-prophet counts are available. Option (a) is less destructive.

**Severity: GATED DEFECT** — D6 directly contradicts D1's explicit written law on homonym handling.

---

### DEFECT 3 — CASCADE CONTAMINATION: D4 and D5 Salih rows derived from homonym-mixed registry

**Scope:** D4 `patience_cooccurrence_layer.json` (Salih row, n=3), D5 `dawah_cooccurrence_layer.json` (Salih row, n=6)

**The problem:** Because D4 and D5 built their co-occurrences by joining sabr/da'wah roots against the D1 occurrence_refs for Salih (73 raw, homonym-mixed), the resulting co-occurrences are contaminated.

| Layer | Salih co-occurrence | Ayah content | Prophet present? |
|---|---|---|---|
| D4 | 18:82 | "wa-kāna abūhu ṣāliḥan" | ❌ Adjective |
| D4 | 28:80 | "wa-'amila ṣāliḥan" | ❌ Adjective |
| D4 | 7:168 | "al-ṣāliḥīn" | ❌ Adjective |
| D5 | 11:62 | "yā Ṣāliḥu qad kunta fīnā" | ✅ Prophet (address) |
| D5 | 18:82 | same as D4 | ❌ Adjective |
| D5 | 35:37 | "wa-man ʿamila ṣāliḥan" | ❌ Adjective |
| D5 | 41:33 | "wa-ʿamila ṣāliḥan" | ❌ Adjective |
| D5 | 46:15 | "wa-ʿamila ṣāliḥan" | ❌ Adjective |
| D5 | 7:189 | "la'in ātaytanā ṣāliḥan" | ❌ Adjective |

D4 Salih row: all 3 entries are the adjective. D5 Salih row: 5/6 entries are the adjective; only 11:62 is the prophet.

**Fix required:** After the referential pass for Salih عليه السلام is complete, re-run the D4 and D5 co-occurrence joins using only the referential-pass-confirmed occurrence_refs. Until then, the Salih rows in D4 and D5 should carry a `"homonym_note"` declaring the provisional status.

**Severity: DOCUMENTED DEFECT** — The co-occurrence data is misleading in its current form. Does not require immediate rebuild but must carry an honest note before the domain seal can be granted.

---

### DEFECT 4 — HONEST-DOCUMENTATION: D2 titled "second ten" but carries 9 prophets

**Scope:** D2 `second_ten_layer.json` (title/subject), D9 `atlas.json` (second_ten: 9), domain-level documentation

**The problem:** The gates audit record for D2 declares `"subject": "the second ten prophets عليهم السلام of Allah ﷻ"`. The D8 canonical sequence lists D2 as "D2-SECOND-TEN". But the layer has 9 rows. Dhul-Kifl عليه السلام — named in the Quran at 21:85 and 38:48 — is absent from all three registry domains.

D1-SOURCES.md acknowledges: *"Dhul-Kifl عليه السلام is a compound name (two tokens). He does not appear in the probe's lemma table because the compound pass must be run separately… His domain slot is not D1."* This declares D1 is not responsible. But the compound pass was never run and no other domain claims him.

The atlas honestly records `second_ten: 9`, but no holding note explains the gap.

**Fix required:** Either (a) add a `"dhul_kifl_status"` note to D2 _meta declaring the compound-name pass is pending and Dhul-Kifl عليه السلام will be inserted upon completion, or (b) rename D2's subject to "second group of prophets — nine registered; Dhul-Kifl عليه السلام pending compound-name pass". The gap is not a content error (his prophethood is a classical scholarly debate) but it must be explicitly documented, not silently absent.

**Severity: DOCUMENTED DEFECT** — honesty law requires the gap to be acknowledged in writing at the domain level.

---

## §4 — CROSS-DOMAIN COHERENCE

| Overlap point | Domain A | Domain B | Agreement? | Notes |
|---|---|---|---|---|
| Ibrahim عليه السلام unique ayat = 63 | D1 unique_ayah_count=63 | D6 ayah_count=63 | ✅ YES | Three-digit anchor, two-domain agreement |
| Nuh عليه السلام = 43 | D1 unique=43 | D6 ayah_count=43 | ✅ YES | |
| Adam عليه السلام = 25 | D1 unique=25 | D6 ayah_count=25 | ✅ YES | |
| Musa عليه السلام unique ayat = 99 | D2 occurrence_refs=99 items | D6 ayah_count=99 | ✅ YES | D2 positions=136 (word-level); D6 correctly uses unique-ayah count |
| Shu'ayb عليه السلام unique ayat = 10 | D2 occurrence_refs=10 items (positions=11) | D6 ayah_count=10 | ✅ YES | D6 correctly uses unique-ayah count; D2's 11th position is a second occurrence within one ayah |
| 'Isa عليه السلام unique ayat = 25 | D3 occurrence_refs=25 | D6 ayah_count=25 | ✅ YES | |
| Zakariyya عليه السلام unique ayat = 6 | D3 occurrence_refs=6 (positions=7) | D6 ayah_count=6 | ✅ YES | |
| Total graph nodes = 24 | D1(10)+D2(9)+D3(5)=24 | D6 prophets=24 | ✅ YES | Atlas also confirms 24 |
| Total graph edges = 527 | Sum of all ayah_counts = 527 | D6 total_edges=527 | ✅ YES | |
| Hud عليه السلام count | D1: word_positions=10, homonym_mixed=true | D6: ayah_count=10, NO flag | ⚠️ CONFLICT — D6 presents raw as clean | Defect 2 |
| Salih عليه السلام count | D1: word_positions=74, unique=73, homonym_mixed=true | D6: ayah_count=73, NO flag | ⚠️ CONFLICT — D6 presents raw as clean | Defect 2 |
| D4 Salih co-occurrences | D1: Salih homonym_mixed=true | D4: Salih row uses raw occurrence_refs | ⚠️ CONTAMINATION | Defect 3 |
| D5 Salih co-occurrences | D1: Salih homonym_mixed=true | D5: Salih row uses raw occurrence_refs | ⚠️ CONTAMINATION | Defect 3 |
| D3 prophet honorifics | D2: "Yusuf 'alayhi as-salam" (correct) | D3: "Yunus" bare (incorrect) | ⚠️ INCONSISTENCY | Defect 1 |
| D6 node honorifics | D2 subset: all carry 'alayhi as-salam | D1+D3 subset: all bare | ⚠️ INCONSISTENCY within D6 | Defect 1 |
| D8 D3 entries | D3: bare names | D8 inherits from D3: same bare names | ⚠️ PROPAGATED | Defect 1 |
| Second-ten count | D2 domain title: "second ten" | D2 actual rows: 9 | ⚠️ MISMATCH | Defect 4 |
| Dhul-Kifl عليه السلام | D1-SOURCES.md: "domain slot is not D1" | No other domain claims him | ⚠️ GAP | Defect 4 |

**Nine out of fifteen overlap points are clean. Six require fixes.** All six map to the four named defects above.

---

## §5 — OPEN-WORK REGISTER (Block 3+)

These are not defects — they are explicitly honest deferred items that the domains correctly declare as future work.

| # | Item | Domain(s) | Block | Evidence of honest declaration |
|---|---|---|---|---|
| 1 | Referential pass for Hud عليه السلام — clean per-prophet count after removing hudan (Jews/guided) occurrences (estimated: 7 clean of 10 raw) | D1, D6 | Block 3 | D1 `homonym_law` field; `block_3_status: AWAITED-REFERENTIAL-PASS` |
| 2 | Referential pass for Salih عليه السلام — clean per-prophet count after removing adjectival ṣāliḥ occurrences (majority of 73 raw are adjective) | D1, D4, D5, D6 | Block 3 | D1 `homonym_law` field; `block_3_status: AWAITED-REFERENTIAL-PASS` |
| 3 | Dhul-Kifl عليه السلام compound-name pass — two-token query required; prophethood status disputed in classical scholarship (documented but not adjudicated) | D2 | Block 3 | D1-SOURCES.md compound-name method note |
| 4 | D7-IN-TAFSIR domain completion — consensus engine has 94 records; only 2 key cross-pillar verses (21:25, 33:40) formally registered; full domain selection pending | D7 | Block 3 | D7 `_meta.law`: "CO-OCCURRENCE ONLY … prophet-verses already documented by the SHELF-SOUND consensus engine"; cache_size=94 declared |
| 5 | D8 geography atlas — all place attributions (Makkah, Ur, Madyan, Miṣr, al-Quds) await named classical sources; currently "AWAITED-CLASSICAL" | D8, D9 | Block 4 | D8 `geography_status` field; D9 `summary.chronology` field |
| 6 | D4 and D5 Salih row rebuild — after referential pass, re-run co-occurrence joins with clean occurrence_refs; current rows are provisional | D4, D5 | Block 3 | Follows from open-work items 1 and 2 |
| 7 | D9 atlas update — rebuild after all source-domain defects fixed and referential passes complete; add post-pass counts for Hud and Salih; include Dhul-Kifl if compound pass confirms prophetic referent | D9 | Block 3+ | Atlas already honest: "aggregated from the sealed domain layers" |
| 8 | SOURCES-SEAL infrastructure — D1-SOURCES.md declares `_C7-D1-SOURCES-SEAL.sha256` "does not yet exist — written by the first run of the domain build script"; same for other domains | All | Block 3 | D1-SOURCES.md §3 |
| 9 | Isrāʾīliyyāt material — the QB-genre (Tha'labi, Kisā'ī) carries substantial genealogical and narrative material flagged ISRA-NARRATIVE; it requires ongoing discipline to keep flagged vs. admitted as confirmed evidence | D1–D3 | Block 3+ | D1-SOURCES.md §6, §7, §9 |

---

## §6 — NINE-SEAT PANEL TABLE

| Seat | Domain | Primary layer | Scale | Key counts | Defects found | Severity | Seal |
|---|---|---|---|---|---|---|---|
| 1 | D1-REGISTRY | `first_ten_layer.json` | 10 rows | Adam=25, Nuh=43, Ibrahim=69/63 | 0 direct; ⚠️ homonym flags present but unfixed in downstream D6 | — | ⚠️ NOT SEALED |
| 2 | D2-SECOND-TEN | `second_ten_layer.json` | 9 rows | Musa=136/99, Harun=20 | 1: domain titled "ten" but has 9 (Defect 4) | DOCUMENTED | ⚠️ NOT SEALED |
| 3 | D3-FINAL-FIVE | `final_five_layer.json` | 5 rows | 'Isa=25, Zakariyya=7/6 | 5: all 5 prophets missing honorifics incl. ﷺ (Defect 1) | GATED | ⚠️ NOT SEALED |
| 4 | D4-PATIENCE | `patience_cooccurrence_layer.json` | 10 rows | patience_rows=10 | 1: Salih row homonym-contaminated (Defect 3) | DOCUMENTED | ⚠️ NOT SEALED |
| 5 | D5-DAWAH | `dawah_cooccurrence_layer.json` | 10 rows | dawah_rows=10 | 1: Salih row 5/6 entries homonym-contaminated (Defect 3) | DOCUMENTED | ⚠️ NOT SEALED |
| 6 | D6-GRAPH | `prophet_graph_layer.json` | 24 nodes, 527 edges | 24 prophets, 527 edges | 2: missing honorifics for 15 nodes (Defect 1); homonym counts unflaged for Hud+Salih (Defect 2) | GATED | ⚠️ NOT SEALED |
| 7 | D7-IN-TAFSIR | `prophets_tafsir_layer.json` | 2 verses, cache=94 | tafsir_verses_cached=2 | 0 defects; domain is scaffolded (honest open-work) | — | ⚠️ NOT SEALED |
| 8 | D8-GEOGRAPHY-CHRONOLOGY | `chronology_layer.json` | 24-prophet sequence | geography AWAITED-CLASSICAL | 1: D3 prophet entries missing honorifics (Defect 1 cascade) | GATED | ⚠️ NOT SEALED |
| 9 | D9-ATLAS | `atlas.json` | aggregate | 24 prophets, 527 edges, rows confirmed | 0 direct; inherits from source domains | — | ⚠️ NOT SEALED |

**Zero domains sealed. Four defect categories. Two categories are GATED (adab law, homonym-contamination law) and must be resolved before any seal proceeds.**

---

## §7 — PILLAR COHERENCE FINDINGS

**Finding 1 — All primary occurrence counts are internally consistent across citing domains** ✅  
Every count shared between D1/D2/D3 and D6 (the graph) agrees. Ibrahim عليه السلام = 63 unique ayat in both. Nuh عليه السلام = 43 in both. Adam عليه السلام = 25 in both. Musa عليه السلام unique ayat = 99 in both. Zakariyya عليه السلام unique ayat = 6 in both. The graph total edges (527) is verified by calculation from the node ayah_counts. The atlas confirms all domain-level aggregates correctly.

**Finding 2 — Adab law violated in D3, D6, and D8** ❌  
D3 carries 5 bare prophet names — four without عليه السلام, one (Muhammad) without ﷺ. This propagates into D6 (graph nodes for all D3 prophets, and for all D1 prophets, are bare) and D8 (D3 prophet entries in the chronology sequence). D2 correctly included honorifics in its gloss field; D1 correctly stores them in its `name_honorific` field. D3 did not. The resulting inconsistency within D6 — where D2 nodes carry honorifics and D1+D3 nodes do not — makes the graph uneven. **Gated defect. FIX-FIRST.**

**Finding 3 — D6 graph propagates homonym-contaminated counts for Hud عليه السلام and Salih عليه السلام** ❌  
D1's explicit HOMONYM-MIXED law prohibits stating any per-prophet occurrence count for Hud or Salih before the referential pass. D6 violates this by assigning `ayah_count=10` to Hud and `ayah_count=73` to Salih in the graph, with no homonym flag, and includes those counts in the graph's total edge count (527). The total edge count is therefore over-counted by at least 3 (Hud's non-prophetic ayat) and potentially by 60+ (Salih's adjective-dominant refs). The honest statement is that the graph's edge count for these two nodes is provisional until the referential pass is complete. **Gated defect. FIX-FIRST.**

**Finding 4 — D4 and D5 Salih rows are cascade-contaminated** ⚠️  
Both the patience and da'wah co-occurrence layers include the Salih row built from the homonym-mixed occurrence list. D4's 3 Salih entries are all adjective uses; D5's 6 Salih entries include 5 adjective uses. This does not produce false positive counts of other prophets — the contamination is isolated to the Salih row. But those rows currently mislead about which ayat actually involve the prophet Salih عليه السلام. **Documented defect. Requires honest-note or rebuild after referential pass.**

**Finding 5 — D2 title-count mismatch is an honest-documentation gap** ⚠️  
Dhul-Kifl عليه السلام is acknowledged in D1-SOURCES.md as a compound-name requiring a separate pass and "not in D1." But no other domain claims him, and D2 is titled "second ten" while having 9 members. The gap is not hidden — the atlas honestly records `second_ten: 9` — but the domain-level documentation does not carry a forward-pointer explaining where Dhul-Kifl will be placed. **Documented defect.**

**Finding 6 — All honest open-work items are consistent across domains** ✅  
The referential-pass deferrals (Hud, Salih), geography deferrals (AWAITED-CLASSICAL), D7 tafsir scope (2 formal + 94 engine cache), SOURCES-SEAL infrastructure, and Isrā'īliyyāt handling flags are consistent across the domains that declare them. No domain overclaims that Block 3 work is done.

---

## FINAL VERDICT

> ## 🔴 FIX-FIRST
>
> **Pillar C7 (The Prophets & Messengers عليهم السلام) — 24 prophets, 9 domains — is internally consistent in all primary occurrence counts, but carries four defect categories that must be resolved before any domain seal.**
>
> **The two gated defects:**
> 1. **Adab** — D3 carries 5 bare prophet mentions (4 without عليه السلام, Muhammad without ﷺ). This propagates into D6 (15 of 24 graph nodes bare, including Muhammad bare) and D8 (D3 entries carry the same bare forms). Fix: add the correct honorific to each affected field in D3, D6, and D8.
> 2. **Homonym integrity** — D6 presents the raw homonym-contaminated counts for Hud عليه السلام (10, includes ≥3 non-prophetic) and Salih عليه السلام (73, adjective-dominant) as clean prophetic-referent `ayah_count` values, without a `homonym_mixed` flag. This directly contradicts D1's declared law. Fix: add `"homonym_mixed": true` and a clear note to both nodes, or remove them from D6 pending the referential pass.
>
> **The two documented defects:**
> 3. **Cascade contamination** — D4 and D5 Salih rows built from the homonym-mixed registry include adjective-only ayat. Fix: add a provisional note now; rebuild from clean refs after the referential pass.
> 4. **D2 title gap** — domain says "second ten" but carries 9 members; Dhul-Kifl عليه السلام has no domain home and no explicit forward-pointer. Fix: add a `dhul_kifl_status` note to D2's _meta.
>
> **Routing once defects are fixed:**
> Fix defects 1–2 first (they are adab/honesty laws, not Block 3 work). Then add provisional notes for defects 3–4. Then run the referential passes (Block 3). Then seal the individual domains in order. Then rebuild D9 atlas. Then run the pillar coherence panel again.
>
> **The open-work register (9 items in §5) is honest and consistent.** No domain overclaims work that has not been done.

---

*OP50 — window closed*