[dillona]

This adds another fully charged causal-history split on exact #7140, improving conditional κ to 0.003643180671645, a gain of 0.000000221652614. Two timed gathers and six later injections isolate helper 1449 at chart 120. Pivot 648 retires there; two explicit owned buffers and one paid rank-seven path split into ranks four and three complete the construction. The productive rank-nine helper splits into ranks five and four. Ali Romman supplied the mechanism, and ToikY's earlier chart-120 lead remains credited.

The native word adds 21 ADDs, two SCALEs, one MOVE, two INITs and two CLEARs. Owned inventory changes from 53 labels/56 lifetimes/peak 44 to 55/58/44. The unchanged 17-stage caller has 110,434 paid positions, 22,342 productive fields, surplus 1,424, zero empty credit, ordinary source-unit work 779,260,044 < 10⁹, and precision c=2362 on the freshly derived 1125/F25 grid. All added storage, scalar, routing and compiler services are charged.

All three direct CPython 3.14.7 commands and the registered isolated CPython 3.12.16 replay passed, matching every frozen receipt field. All 24 scientific hashes remained unchanged. Coverage includes the exact timed certificate; whole arbitrary-input forward/adjoint; COPY and exact lifecycle checks; 18 native corruption controls; six sets of 512 chart/phase columns; finite physical routes; all 3,038 paid/helper maps and 104 main boundaries; both directed prices; fresh owned-stock, ordinary, precision and fixed-tape costs; and 54 strict + 4 functional outer conditions. The lifecycle assertion remains enforced against (58,44,55), including rejection of an extra uncharged lifetime.

git fetch --no-tags https://github.com/CrocSwap/integer-mult-bounds.git main
python3.12 -B scripts/research_package.py --package research/p104-temporal-chart120-split

INPUTS.json retains the complete 30-file #7140 package at a95783a8c861248a759136d4285ea725499cf231, including its nested closure and notices. Six implementation files remain byte-identical and are authenticated through REUSE.json. Candidate native SHA256: c38b79279f665d46d795501de6645a8a16d9324c22199fa615313b6abfdc3e2f. Discovery stopped at its first positive after 715 of 7,667 individually affordable paths; the remaining domain is unprocessed. The invoice-specific crossing and BIT headroom are documented; no temporal optimum is claimed.

Credit Ali Romman, Jerry Chen/CubeJerry, Matarener, Rohan Arun (#7113, #7122, #7124, #7131), ToikY, Cinna Davis/1337isnot1337, Dugongue, DreamingOfClouds, Douglas Colkitt/#7000 reviewers, the dillona #7140 predecessor and every retained source author. Prepared for dillona with substantial OpenAI Codex assistance; no subagents. Apache-2.0.

The five inherited assumptions (native_all_columns, owned_erasure, compiler, precision_tape, analytic_recovery), independent human/formal acceptance, maintainer admission and compiled top-level execution remain open. The selected frontier is unchanged. Hosted CI status is separate from local scientific replay.

Task p104-side-correction; changed exact source and timed access contract are recorded in TASK.json. Canonical e545235 refetched unchanged at 00:20 UTC; revision 8 / selected #7000 unchanged. Complete snapshot: 442 open PRs at 00:21:17 UTC and 7055 Discussion records at 00:22:18 UTC. Sparse local coverage: 12 working records, 10178 HEAD note paths, 10167 unsearched. New Rohan Arun draft #7142 at d446d89 reports provisional kappa .003643216822476 on the same #7140 source, with full verification pending. It uses an endpoint-funded T2+1,T7+1,T8-1/P1+1 change, distinct from this temporal T4+1,T5+1,T9-1/P3+1,P4+1,P7-1. This package does not claim to beat that draft or compose its change. #7141 covers separate fixed-profile two-blocker searches. Exact source7140 temporal scope was claimed before discovery.

Open next questions: independently replay the timed e1449 equations and all owned/caller costs; then, after a new upstream/overlap checkpoint, test an uncovered fully priced temporal domain. No global barrier follows from this first-positive search.

Canonical investigation and coverage record.

Concurrent work: Rohan Arun’s draft #7142 has a higher provisional κ=.003643216822476 from a distinct endpoint-funded change on #7140. This PR does not claim to beat or verify that draft. Possible composition remains open and requires a new complete replay.

Final canonical main fetch at 00:31 UTC remained e545235; newest-eight PR and exact #7142 status rechecks were unchanged at 00:30 UTC.

Hosted status at public head 9ff298b: run42932212575 failed in onboarding:760tests,onefailure,24skipped. CurrentCiNavigationTests.test_current_ci_section_matches_selected_record expects make frontier7000-verify; unchanged CI navigation still references#6030. The scientific replay job was skipped. This is separate from the passing direct and registered local scientific replay; no hosted scientific PASS is claimed.