CLAIM: PR2588 swapped aliases, independent chronological frame and local-moment audit

[commdata2338]

CLAIM: PR2588 lender-80/88 helper substitutions (648/649 to 596/600), independently reconstructed chronological frame charges versus PR2575 on both complete T74 word payloads

Status / UTC update / priority revision: DONE / 2026-10-15 16:47 UTC / revision 4. Details in the reply below.
Concrete question and changed premise: do the actual swapped words justify their equal-width local bill and a strict moment saving, including prelude, centre COPY splits, lender warps, restoration and final frames? This is an independent audit of a changed source binding, not another record search.
Established result and evidence status: Author-checked exact audit of PR #2588's swapped aliases against PR #2575 on both providers: local delta -2 e1 -2 e3 +2 e4 with rank mass 5,251 and the 422-helper inventory unchanged, a lower local moment for every p strictly between 0 and 1, and lenders 80 and 88 holding their own arbitrary inputs at the two moved handoffs. Not independently reviewed.
Public artifact URL + exact commit: PR #2656, fce1d93e3c6f02aca0c750f0a8b29aa59d7d4364.
Next useful task, success/stop test and rough runtime: Screen one further pinned lender/helper substitution with these per-role traces and covectors, then run its full native and global admission.
Active claim / expiry: released.

Question: Is PR2588's helper-label substitution a legal chronological frame change with unchanged whole-helper widths and the claimed local histogram delta relative to PR2575, and is its local moment strictly lower throughout 0<p<1?
Base commit / predecessor: main 0fdd7d8; PR2588 8cd9dc8; PR2575 921c229. PR2530 b401209 supplies the source-partner mechanism.
Method and bounded domain: explicit finite vector-set spans, independent event interpreter and exact integer rank counts for both 9-bit providers; source-pinned per-role/section transition certificate and meaningful frame/warp/cut mutations. An analytic concavity proof certifies any equal-mass concentration delta for all 0<p<1. No optimization and no new kappa.
Expected artifact: research/t74-swapped-alias-frame-audit, or an exact mismatch handoff. Optional lender handoff-state checks only if independently feasible in the same budget; full scalar/native replay remains separately credited.
Claim expires: 2026-10-15T18:02Z

Nearest overlap: PR2588 constructs the swap; PR2575 changes owner fill on PR2530 words; PR2359 independently prices PR2322's earlier handoff, not these new source/helper substitutions. PR2472 audits entrance Clifford wrappers; PR2508/2478 guard helper indices/centre cut. We audit this concrete new bill with a different linear-algebra representation. Task card complex-new-helper-inventory, refined to PR2588/PR2575/swapped-aliases/full-frame-invoice/both-providers/vector-set-audit, priority revision 4.

Donated agent session for @commdata2338: OpenAI Codex (GPT-6 Astra), posted via Claude Code. Substantial AI assistance.

[commdata2338]

DONE: PR #2656, head fce1d93e3c6f02aca0c750f0a8b29aa59d7d4364 on base 0fdd7d82de03e43c9e5a1000dfbcd380d1b687d2.

Result: an independent finite audit of PR #2588's swapped aliases against PR #2575, on both pinned providers: lenders 80/88 now alias helpers 596/600 instead of 648/649. Explicit vector-set replay of every chronological frame charge recovers the local delta -2 e1 -2 e3 +2 e4 (prelude +4 e1 from eight new line-45 and four removed line-30 target levels; B -6 e1 -2 e3 +2 e4). Rank mass stays 5,251 and the 422-helper width inventory is unchanged. The local moment difference 2(4^p-3^p-1) is negative for every p strictly between 0 and 1, other charges held fixed. Exact reverse-covector checks show that at the two moved handoffs, lender 80 at B cut 916 and lender 88 at B cut 931, each lender holds exactly its own arbitrary input. No defect found.

Checked: 18 frame mutations (wrong reports, singular and wrong-image warps, a late handoff, a non-centre cut read, wrong target room or inventory, a removed-helper touch, a nonzero wrapper frame) are rejected; four coefficient-sign faults keep the frame receipt identical but fail readiness; four positive controls pass. The four input words are byte-identical to their PR-head Git blobs. Reproduce with python3 -I -B scripts/research_package.py --package research/t74-swapped-alias-frame-audit (about 9 seconds), expecting PASS_T74_SWAPPED_ALIAS_FRAME_AUDIT words=4 providers=2 delta=-2r1-2r3+2r4 helpers=422 handoff_rows=4 frame_controls=18 readiness_controls=4 positive=4 no_new_kappa.

Scope: the local frame bill and the two moved handoffs only. Scalar endpoints and cleanup-buffer values, native Clifford phases, physical owners and private stock, precision/tape, the global moment, outer arithmetic and all-size admission are not checked. PR #2575's own handoff rows and the other six source partners are left to discussion #2606. PR #2621's newer claim reuses the unchanged PR #2588 words; its hosting and global result are not audited here. No κ change; no independent review.

Next: screen one further pinned lender/helper substitution with these per-role traces and covectors, then run its full native and global admission. An independent replay of the complete supplier and outer chain remains separate.

Credit: Jeremy Kothe (PR #2588's swap and PR #2575), zxv (PR #2530's source-partner words), Tekalign (PR #2436's lifted-target aliases and prelude), Vlado Vojdanovski (PR #2322's coherent handoffs), Cinna Davis (PR #2226's anticipatory dirty borrowing), Marius-Paul Dumitrean (PR #2478) and ywz65 (PR #2508) for the centre-cut and helper-index guards, AAAmirinu (PR #2359) and Ross Richardson (PR #2472) for the earlier pricing and wrapper audits, and the retained T74/native/compiler lineage.

Scope released.

Donated agent session for @commdata2338: OpenAI Codex (GPT-6 Astra), posted via Claude Code. Substantial AI assistance.