# Adversarial review log — addiction-brain-disease (revised edition)

Authoring model: **Sol 6.1**. Adversarial reviewer: **Claude Opus 5.5 (Anthropic)**. Date: **2026-10-14**.
Round 2; group 12 of `NEW_REVIEW_INVENTORY_2026-10-14.json`. Canonical copy is the 9 October 2026 revised edition.

## Summary

| Item | Value |
|---|---|
| Level before | Γ — Mechanistically plausible, partial system support / Moderate |
| Level after | Γ — Mechanistically plausible, partial system support / Moderate (unchanged) |
| Findings: Critical | 0 |
| Findings: Major | 2 |
| Findings: Minor | 3 |
| Findings: Editorial | 1 |
| Rejected findings | 9 |
| PMIDs checked / failed | 56 inventory + 3 supplemental / 0 |
| Load-bearing claims spot-checked | 12 |
| Validator | PASS (0 fail, 0 warn, 30 pass) — identical before and after. |
| Approval | **Approved as revised.** |

## Findings

Full table (AR-1 to AR-6), the nine rejected findings and the residual limitations are in the `## ADVERSARIAL REVIEW AND APPROVAL` section of `addiction-brain-disease.review.md`. The three revision-introduced regressions, which matter more than inherited problems:

- **AR-1 (Major).** This is the known systematic artifact, and this case has it. The 9 October integration spliced in a "Consensus consistency validation" section documenting a raw Consensus Pro and browser response, connector-search discovery leads and cited Consensus markers, while leaving the preserved REVIEWER NOTES sentence stating that no Consensus work was performed. Repaired by scoping the original sentence to the original 2026-10-09 review pass and recording that the dated validation is a later, separately scoped supplement that changed no grade and is not blind calibration either. Neither block was deleted.
- **AR-2 (Major).** The validation block's closing paragraph asserts that "the earlier approval does not certify this supplement", presupposing an earlier adversarial approval. Verified against `hsef_v53_remaining_priority_01_05_2026-10-09/addiction-brain-disease/addiction-brain-disease.review.md`, which has no approval section or banner. Rescoped in both the review and `consensus-addendum.md`.
- **AR-3 (Minor).** The "Primary-source checks" table was broken by the revision: the row added for the Rewarding Recovery trial sat behind two blank lines and rendered as a detached header-less table, and three rows were missing their trailing cell delimiter. Structure repaired; content unchanged.

Remaining: AR-4 confidence rationale rewritten off access depth (band stands at Moderate under the lead two-step test, since the review documents real evidence-quality reservations); AR-5 Divergence artifact added; AR-6 one empty `DOI: .` field and two fused `PMIDs...` labels repaired.

## Source-level verification of the revision's own additions

Because the four source-level rows were added by the 9 October revision, each was checked line by line against PubMed. All four are accurate, including the figures most liable to drift: 67 and 62 smoking cohorts with an independent 186-patient alcohol-risk cohort and the authors' own "candidate target" framing (35697842); 68 patients and 42 controls with 27 abstinent and 33 relapsed at six months (37567334); the 65-person Brazilian trial randomized 33 and 32 with the denominator distinction preserved (27442691); and 114 American Indian adults in a 2x2 factorial trial with alcohol-abstinent odds ratios of 2.4 to 4.8 and drug contrasts losing significance under multiple imputation, reported with the abstract's own word "inconclusive" (33220122). The supplemental sources are correctly excluded from the 56-source inventory and the double-counting warning is correct.

## Strings for the lead to propagate to package-level files

None. No level, confidence, role, momentum value, source count or claim count changed, so no package-level scoring string needs updating. The earlier edition at `hsef_v53_remaining_priority_01_05_2026-10-09/addiction-brain-disease/` is a divergent copy and was **not** overwritten.

**Approval:** Following adversarial review and the corrections listed above, the revised document is approved. — Claude Opus 5.5 (Anthropic), 2026-10-14.