[firejune]

Change

The config producer's generated skirt fails the unchanged setup seam. Adopt the selected explicit mesh at parts 1.64.0, retain its turn depth, and freeze 31 producer files with byte and decoded-RGBA verification. Archive the former keeper so historical v1/v3 replay remains source-bound.

Separately add opt-in accepted-dance/v4: only the bottomwear attachment changes from v3; literal motion bytes and every other rig subtree remain identical. V3 stays the default pending owner eye acceptance. This is the two-delivery scope recorded in #71; the producer lacks the accepted dance's alternate/control bones, so its dance metrics are inapplicable.

Evidence

Verification	Exit / result
Same parts 1.64.0 baseline vs explicit producer	baseline exit 1: seam mean .962, 5052 pixels over 40; candidate exit 0: mean 0, 0 pixels over 40
Independent fresh producer build / inventory	exit 0; all 31 files byte-identical, PNG dimensions/RGBA identical
Producer check	11/12 bars measured, all measured bars pass; STRESS_OFF_CHAIN skipped, CLOTH_SETTLES warning
Native Spine 4.3.26 roundtrip on fresh packed target	overall exit 0, import/export/validate/diff/render-check exit 0; summary step has no subprocess; rewrites recorded
V3/v4 released core 2.67.2 gate	each 23 PASS, 0 FAIL, 13 SKIP, 14 excluded
Fixed viewport original eval.py	v3 left/right 31/13 → v4 20/4; zero bottomwear fold frames in both
V3 exact replay	all three compiled hashes and 391-frame native concatenation match accepted record
Final targeted checks	inventory positive/byte/pixel/missing/empty plants pass; v1/v3 preflights and v4 real compile-reader preflight exit 0; producer stored inventory exit 0; diff/hygiene clean

Bones remain identical by name in the producer; raw order swaps two turn bones and the matching transform bone list. Physics and motion remain identical. This explicit scope amendment replaces the initial raw-array identity requirement. Native editor rewrites are reported, not claimed byte identity; macOS does not provide the Linux descendant-boundary proof, and no Spine processes remain.

Known shins folds remain 269 frames. V4 silhouette deviation against accepted v3 is 15.425046px max (including 14.164224px setup shift); the earlier 4.6px figure used another baseline. The mesh reduces the stair artifact, without claiming elimination.

The owner video was delivered privately: 51 frames f240–290, full figure and both flanks at 4x nearest, 3072x1600 H.264, SAR 1:1, 1,367,689 bytes. Its receipt and small comparison sheet are tracked; 782 native PNGs were hash-checked and deleted after sheet review. Owner verdict and HQ exact-head sign remain merge gates. No renderer patch, track drop, control alias, family tool change or private measurement scripts are included.

Closes #71.

@codex review

[chatgpt-codex-connector]

Codex Review Summary

This comment shows the latest Codex review activity on this pull request.

Review	Status	Commit	Review trigger
📝 Code Review	✅ Completed 2026-10-25T08:22:03.232411Z	8dd377f	Manual request

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.

[chatgpt-codex-connector]

Codex Review: Didn't find any major issues. Hooray!

Reviewed commit: 8dd377fbf2

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[firejune]

LGTM #72 at 8dd377fbf28d0fb26a93be37a70e614a8d77fb75. Owner eye check on the v3 | v4 video (f240–290, full figure plus both flanks at 4×), quoted: "v4 looks good, merge it." CI is green and the one AI review found nothing. v4 drops the hem-share staircase from 31/13 to 20/4, with 0 bottomwear fold frames.