# ADVERSARIAL PANEL REPORT — P30 Wing B: Revealed-Meaning Machinery
**Window:** SP3 (Opus)  
**Scope:** D-B1-MARFU · D-B3-OCCASIONS · D-B4-SIRAH-SYNC · D-B8-DOSSIERS · strata.py  
**Date:** 2026-09-03  
**Verdict:** FIX-FIRST (3 findings; 5 checks pass clean)

---

## METHODOLOGY

Attack executed across five vectors per mission brief. Every finding is anchored to an exact `file:field:value`. A clean pass was earned on most checks; three findings require correction before the Wing B machinery is declared coherent.

---

## FINDINGS REQUIRING FIX

### FINDING 1 — ERA-STAMP ERROR: Al-Amali-Saduq:272 (marfu' content stamped "successor")
**Severity:** HIGH  
**File:** `D-B1-MARFU/marfu_layer.json`  
**Entries affected:** verses 1:1, 1:2, 1:3, 1:4, 1:6 — collection `Al-Amali-Saduq`, id `272`  

**The law (strata.py, line 5):**  
> "every meaning-claim is stamped with the era of its EARLIEST transmission"

**What the entries say:**  
All five entries carry `era_stamp: "successor"`. The partial quote_en for verse 1:2 reads:

> "…from Amir al-Mu'minin (as). He said: **The Messenger of Allah (s) said: Allah said: I have divided the Fatiha of the Book between Me and My servant.** Half of it is for Me, and half of it is for M…"

The **ultimate speaker is the Prophet ﷺ** conveying a hadith qudsi. Amir al-Mu'minin Ali رضي الله عنه (a Companion) reports it from the Prophet ﷺ. Under strata.py the **era of earliest transmission is prophetic** (the Prophet ﷺ is the speaker).

**Why the integrator chose "successor":** The chain passes through post-Companion Imams (al-Hasan al-'Askari → forefathers → Ali → Prophet ﷺ). The integrator appears to have stamped the **proximate identified link** in the Shia chain (which is post-Tabi'i) rather than the **ultimate attributed speaker** (the Prophet ﷺ).

**Why this is wrong under the law:** strata.py stamp_report keys on the speaker, not the intermediate chain. The identical report through Sunni chains (malik:20, nasai:34) is correctly stamped "prophetic." The same content (the divided-prayer/Fatiha qudsi hadith) coming through the ahl al-bayt chain should also be "prophetic" — the division point is the CONTENT, not which collection carries it.

**Consequence for dossiers:** For verse 1:1, the dossier's `transmitted_meaning_oldest_first` opens with five "successor" entries. If Al-Amali:272 is correctly re-stamped "prophetic", verse 1:1's dossier would gain a prophetic-era entry it currently lacks — a material change to the oldest-first rendering.

**Required fix:** Re-examine the full hadith text (not the truncated quote) for each Al-Amali:272 entry. If the ultimate attributed speaker is the Prophet ﷺ, restamp to "prophetic." If the Imam's own words (not a prophetic attribution) are what constitute the explanation, document that reasoning explicitly in the integrator_note and confirm "successor" is the correct classification.

---

### FINDING 2 — B3 CLASSIFIER MISCLASSIFICATION: abudawud:3001 tagged "context-only"
**Severity:** MEDIUM  
**File:** `D-B3-OCCASIONS/occasions_layer.json`  
**Entry:** verse `3:123`, source `abudawud:3001`, `class: "context-only"`, `markers_found: []`

**Meta claim:** "statement-detection certified" (D-B3 _meta.method)

**What the entry's quote_en says:**

> "You have never met people like us." **Allah Most High revealed about this the following verse** "Say to those who reject faith, soon will ye be vanished… one army was fighting in the cause of Allaah, the other resisting Allaah."

The phrase **"Allah Most High revealed about this the following verse"** is an occasion statement — the speaker is asserting a specific revelation occasion. The classifier's marker vocabulary should have caught a construction of this form. It did not. `markers_found` is empty and `class` is "context-only."

**The correct class is "occasion-stated."**

**Note on verse_binding:** The verse cited in the quote ("Say to those who reject faith") is not 3:123. `verse_binding: "claimed-requires-check"` is already on this entry, which is the correct safeguard against feeding this into a dossier unchecked. The misclassification is therefore isolated: a missed "occasion-stated" label, not a contaminated dossier. But it means the layer's occasion count (238 occasion-stated of 414) is understated by at least 1, and the "certified" claim is falsified by this counter-example.

**Required fix:** Review the classifier's marker set for "revealed about this," "revealed concerning this," and "was revealed about this" — forms with a demonstrative pronoun as object rather than a noun phrase. Rerun the classifier and recheck the occasion count.

---

### FINDING 3 — 'AS IT WAS REVEALED' OVERCLAIM in D-B8 meta
**Severity:** MEDIUM  
**File:** `D-B8-DOSSIERS/dossiers_v1.json`  
**Field:** `_meta.layer: "D-B8 dossiers v1 - 'as it was revealed', assembled"`

**What the dossier for verse 1:1 actually contains:**

```
transmitted_meaning_oldest_first:
  [1] era_stamp: "successor"  → Al-Amali-Saduq:272
  [2] era_stamp: "successor"  → Al-Kafi-Volume-3-Kulayni:1873
  [3] era_stamp: "successor"  → Uyun-akhbar-al-Rida-Volume-1-Saduq:263
  [4] era_stamp: "successor"  → Uyun-akhbar-al-Rida-Volume-2-Saduq:391
  [5] era_stamp: "successor"  → Uyun-akhbar-al-Rida-Volume-2-Saduq:460
  [6] era_stamp: "classical"  → Man-La-Yahduruh-al-Faqih-Volume-2-Saduq:1570
```

The oldest available transmission for verse 1:1 is **successor-era**. There is no prophetic-era or companion-era explanation of the basmalah in the current harvests. The absent_note correctly says "strata and layers not listed are ABSENT in the current harvests — honest gaps." But the **_meta.layer title** still claims "'as it was revealed.'"

**Why this matters:** "As it was revealed" implies the dossier delivers the meaning contemporaneous with revelation. It does not. It delivers "as transmitted from the earliest available era in the current corpus." For verse 1:1, that era is the 3rd–4th century AH Shia collections. The claim is unearned — not because the work is bad, but because the corpus gap for 1:1 is real and the title should not paper over it.

**strata.py is honest** (docstring: "every meaning-claim is stamped with the era of its EARLIEST transmission"). The dossier builder's title appropriated the stronger rhetoric.

**Required fix:** Change `_meta.layer` to: `"D-B8 dossiers v1 - transmitted meanings oldest-first, assembled"`. The 'as it was revealed' language belongs in the aspirational framing document (strata.py), not in the machine-output meta of a specific-corpus run where the prophetic stratum is absent for multiple verses.

---

## CHECKS THAT PASSED CLEAN

### CHECK A — Era stamps: 6 B1 entries spot-verified

| Entry | Verse | Collection | Class | Stamp | Speaker | Verdict |
|---|---|---|---|---|---|---|
| ibnmajah:89 | 113:5 | Sunni | usage-witness | prophetic | Prophet ﷺ recites ruqyah | ✅ CORRECT |
| ibnmajah:31 | 112:3-4 | Sunni | usage-witness | prophetic | Prophet ﷺ affirms greatest-name supplication | ✅ CORRECT |
| Al-Tawhid-Saduq:80 | 112:2-4 | Shia | explanation | companion | al-Husayn ibn Ali رضي الله عنهما answers Basra — his own words, a Companion | ✅ CORRECT |
| malik:20 | 1:2-1:7 | Sunni | explanation | prophetic | "I heard the Messenger of Allah say that Allah the Blessed divided prayer…" — marfu' | ✅ CORRECT |
| Al-Amali-Saduq:272 | 1:1-1:6 | Shia | explanation | **successor** | Content says Prophet ﷺ said it — **see FINDING 1** | ❌ FLAGGED |
| Al-Kafi-3:1873 | 1:1-1:3 | Shia | explanation | successor | Imam's own transmitted explanation (full text needed to confirm content is maqtu', not marfu') | ⚠️ CONDITIONAL |

Conditional note on Al-Kafi-3:1873: the visible truncated quote shows a polemical framing ("What does the enemy narrate?… They have spoken a lie") before reaching the Fatihah explanation proper. The class="explanation" and era_stamp="successor" may be correct if the actual explanation is the Imam's own words — but the integrator_note should distinguish the embedded polemic from the tafsir content. This is flagged for panel attention, not counted as an error pending full-quote verification.

---

### CHECK B — Oldest-first law: all 15 dossiers

Every dossier in dossiers_v1.json was checked for era ordering across its `transmitted_meaning_oldest_first` array. The strata rank is: prophetic(0) < companion(1) < successor(2) < classical(3) < modern(4).

Results:

| Verse | Oldest stamp | Ordering | Verdict |
|---|---|---|---|
| 1:1 | successor | successor→classical | ✅ |
| 1:2 | prophetic | prophetic→successor→classical | ✅ |
| 1:3 | prophetic | prophetic→successor(×5)→classical | ✅ |
| 1:4 | prophetic | prophetic→successor(×3) | ✅ |
| 1:5 | prophetic | prophetic(×2) | ✅ |
| 1:6 | prophetic | prophetic→successor | ✅ |
| 1:7 | prophetic | prophetic(×2) | ✅ |
| 33:33 | companion | companion(×1) | ✅ |
| 55:2 | successor | successor(×2)→classical | ✅ |
| 55:13 | — | empty | ✅ |
| 55:46 | successor | successor(×1) | ✅ |
| 112:2 | companion | companion(×1) | ✅ |
| 112:3 | prophetic | prophetic→companion→successor(×2) | ✅ |
| 112:4 | prophetic | prophetic→companion→successor(×2) | ✅ |
| 113:5 | prophetic | prophetic(×1) | ✅ |

**OLDEST-FIRST LAW: ALL 15 DOSSIERS PASS ✅**

---

### CHECK C — B3: 6 occasion-stated rows sampled

| Source | Verse | Markers found | Quote text | Truly states an occasion? |
|---|---|---|---|---|
| abudawud:2512 | 9:100 | "was revealed about" | "This verse was revealed about us, the group of the Ansar" | ✅ YES — direct, first-person |
| abudawud:3971 | 3:123 | "was revealed about" | "And no Prophet could (ever) be false to his trust was revealed about a red velvet" | ✅ YES — clear occasion |
| ibnmajah:1822 | 9:100 | "was revealed about" | "the following was revealed about the one who did that" | ✅ YES — clear |
| ibnmajah:2986 | 9:100 | "revealed concerning" | "this was revealed concerning some people among the Ansar" | ✅ YES — clear |
| ibnmajah:15 | 9:100 | "revealed concerning" | "I think this verse was revealed concerning this matter" | ⚠️ SOFT — speaker hedges with "I think"; technically an occasion claim but conditional |
| abudawud:2740 | 3:123 | "was revealed about" | "I thought something was revealed about me owing to my speech" | ⚠️ SOFT — reported thought, not a direct revelation claim |

4 of 6 clearly state an occasion. 2 are soft hedged claims correctly admitted under the classifier's marker rule (the marker IS present), but both carry `verse_binding: "claimed-requires-check"` which is the appropriate safeguard. No hard error.

**B3 OCCASION TRUTH CHECK: PASS (with the classifier misclassification at abudawud:3001 flagged separately as FINDING 2) ✅**

---

### CHECK D — claimed-requires-check rows in dossiers

All `occasions_cell_bound` entries in all 15 dossiers were checked against the occasions layer. The rule: rows with `verse_binding: "claimed-requires-check"` MUST NOT appear in a dossier.

Findings:
- Dossiers 1:1 through 1:7, 55:2, 55:13, 55:46, 112:2–4, 113:5: `occasions_cell_bound = []` — no occasions from the hadith layer used ✅
- Dossier 33:33: `occasions_cell_bound` contains three entries:
  - `tafsir_shards/33.json:33:33:tabari`
  - `tafsir_shards/33.json:33:33:miqbas`
  - `tafsir_shards/33.json:33:33:wasit`

All three are tafsir shard rows. Shard rows are cell-bound by construction (the shard cell IS the verse under discussion); they do not carry `verse_binding: "claimed-requires-check"`. This is correct.

**NO claimed-requires-check rows appear in any dossier. ✅**

---

### CHECK E — One-ummah law: Shia/Sunni adjudication scan

Four candidate sites examined:

**33:33 dossier:** The single transmitted-meaning entry (Kitab-al-Ghayba-Numani:30) is class="deferred-to-panel." The occasions from Sunni tafsir (tabari, miqbas, wasit) are in `occasions_cell_bound`. Neither side's position is declared correct. The difference is preserved (`differences_preserved: 1`). ✅ No adjudication.

**55:2 and 55:46:** All entries from Shia collections are class="deferred-to-panel." No Sunni interpretation of these verses appears in the dossier (gaps are honest: absent_note). No adjudication by presence or absence of verdict. ✅

**112:2–4:** Shia sources (Al-Tawhid-Saduq, Al-Kafi) and Sunni sources (ibnmajah) appear in the same dossier, ordered by era without any ranking verdict. The integrator_note for ibnmajah:31 says "usage, not explanation"; the integrator_note for Al-Tawhid-Saduq:80 says "explanation." These class distinctions are made on content grounds, not tradition grounds. ✅ No adjudication.

**Al-Kafi-3:1873 embedded polemic (flagged for attention, not counted as error):** The quote_en for this entry opens with "What does the enemy narrate?… They have spoken a lie" before reaching the Fatihah explanation. This is the Imam's reported critique of Sunni prayer forms. The layer does NOT add its own verdict — it carries the text. However the class="explanation" label without any polemic flag risks presenting the polemical framing as straightforward Fatihah tafsir. Panel should decide: add a `polemic_flag` field to this entry, or split the explanation content from the polemical frame. This is an editorial decision, not a law violation in the current data.

**ONE-UMMAH LAW: NO ADJUDICATION FOUND. ✅ (Al-Kafi-3:1873 polemic frame flagged for editorial decision)**

---

## SUMMARY TABLE

| # | Attack | File | Verdict |
|---|---|---|---|
| F1 | Era stamp: Al-Amali-Saduq:272 stamped successor, content is marfu' | D-B1-MARFU/marfu_layer.json | ❌ FIX |
| F2 | Classifier missed occasion at abudawud:3001 — context-only, should be occasion-stated | D-B3-OCCASIONS/occasions_layer.json | ❌ FIX |
| F3 | 'As it was revealed' meta title overclaims when prophetic stratum is absent | D-B8-DOSSIERS/dossiers_v1.json | ❌ FIX |
| A | Era stamps: 4 verified correct, 1 flagged (F1), 1 conditional | D-B1-MARFU | ⚠️ CONDITIONAL |
| B | Oldest-first ordering: all 15 dossiers | D-B8-DOSSIERS | ✅ PASS |
| C | 6 occasion rows sampled: 4 clear, 2 soft-but-admissible | D-B3-OCCASIONS | ✅ PASS |
| D | claimed-requires-check rows in dossiers: zero | D-B8-DOSSIERS | ✅ PASS |
| E | One-ummah law: no adjudication found | All Wing B layers | ✅ PASS |

---

## PANEL VERDICT: FIX-FIRST

Three fixes required before Wing B is declared PILLAR-COHERENT:

1. **Resolve the Al-Amali-Saduq:272 era-stamp question** (marfu' content vs. ahl al-bayt chain era) with a documented decision — either re-stamp to "prophetic" with reasoning, or confirm "successor" with an explicit integrator note explaining why the chain era overrides the content attribution. The current state carries an undocumented judgment that contradicts the strata.py law on its face.

2. **Fix the abudawud:3001 misclassification** and audit the classifier's marker set for demonstrative-pronoun occasion forms.

3. **Retitle the D-B8 meta** from "'as it was revealed'" to "'as transmitted, oldest-first'" — the honest description of what the layer delivers when the prophetic stratum is absent from the corpus for a given verse.

The oldest-first ordering law (the structural spine of the wing) holds across all 15 dossiers. The one-ummah law holds. The claimed-requires-check firewall holds. The foundation is sound; the three findings are correctible without architectural change.

---
*SP3 adversarial panel — read-only — no dossier or layer file was modified.*