# Disposition record — One-Direction Obstruction (final version: v1.1.3)

## Pass 1 — self-audit, rounds 26–30 (target: v1.1)
All quantitative claims CONFIRMED by the Verifier (7/7 rows on Thm 1; all rows on CF1 / frob / def:standard / rem:true). Findings were definitional-precision only. Applied in v1.1.1: Def 2.2 gains unramified + fixed-signature clause; X_Eis ⊂ C^d defined explicitly with "local complex dimension" defined; five-representation exhaustiveness softened to covers/including + we-are-not-aware; Thm 1 proof's vague meromorphic-extension sentence replaced by the precise chain (countable union of lines → Lebesgue-null → empty interior → no ≥2-dim analytic germ → transverse partials undefined); Λ_F identified as the unit-log image. Classes: 4× VERIFIED-FIX (minor), 2× ACKNOWLEDGED-SCOPE, 0× REFUTED-against-the-math. Export bug (rounds 1–25 duplicated from a prior session under this paper's header) logged as CQ evidence B4/B6; this package's extract is the genuine content only.

## Pass 2 — verify pass, rounds 1–5 (target: v1.1.2)
Scores: units 82 (was 78), CF1 88 (was 72), frob 72, def:standard 72 (was 68), rem:true 68. Fix verification: all v1.1.1 edits confirmed. New findings, applied in v1.1.3: (R3 GAP 1, VERIFIED-FIX, theorem statement, Matt-approved) thm:frob scheme class made explicit — M integral, flat over Z, every fiber integral of pure dimension n ≥ 1 — closing the jumping-dimension/reducible-fiber loophole; (R5 GAPs 1–4, VERIFIED-FIX, cosmetic) rem:true gains the Tate+Bump citation, states rank Λ_F^∨ = d−1, cross-references the Thm 1 closing chain, and disambiguates the full unramified family vs the admissible slice. Coverage note: §6.3 and Appendix A were not machine-covered in either pass (target selection follows the paper's theorem environments — CQ backlog C6); they were covered by the external lock review instead.

## Lock review — third model family (target: v1.1.3)
All nine targets VERIFIED or ACKNOWLEDGED-SCOPE-by-design. "No load-bearing fix needed before Zenodo." §6.3: "correct, no overclaim." Methodological note: "matches FAQ policy." APPROVED FOR CASCADE. Adopted by M. Goss as the G1 sign-off, Aug 28, 2026.

## Standing verdict
Three model families — builder/self-audit, cross-audit (V14 foundation), lock review — concordant on all mathematics. Across the entire process: two theorem-statement-level edits, both hypothesis-narrowing; zero refuted mathematical claims in any pass.