[Clear1nsight]

Change and mathematical scope

The fixed, completed 112-port complex five-FULL/one-THIN profile cannot fund the optimistic target cap for any 33/10000 <= C < 1, under the declared independent-stage, unimodular-boundary and canonical odd-Q monomial-output hypotheses. This package makes that narrow negative result reproducible. It supplies no new κ, caller or universal six-query barrier.

At C = .0033, the existing rational witness gives the exact upper bound
G = -668754569884463/7968172621500 < -41541/500 < 0.
G bounds the signed power moment from above; it is not the moment's exact value. Concavity caps four 112-field main banks, whose nonnegative integer stage degrees total ten, by 224H17 + 224H25. Signed-rank separation extends negativity analytically across the interval without a power/log/exp engine or case enumeration.

The actual eight-H8 neutral completion is fixed in this profile. It is not a universal minimum. The author's BIT n4/1920 class, current BC120/global120 callers, different completions, coded outputs, arbitrary six-query controllers and global κ remain open. No new productive target, physical embedding, compiler/precision/stock funding or outer transfer is admitted.

Task and overlap

This is a changed, explicitly assigned independent-moment support task on canonical 1a3ced60…. The sole all-author monitor confirmed unchanged canonical main, selected record and guidance through 11:37:49 UTC, with no overlapping fixed, completed 112-port complex five-FULL/one-THIN ceiling observed. The construction owner cleared this separate support contribution. Science commit c3fdfe2934aafdce0458fbe5cf0c6e1542978541 directly descends from canonical 1a3ced60…. Publication head a72db1fdffc9532b075f3706ed2c6e8654d6c978 adds only the unchanged, previously reviewed CI-navigation documentation repair from #5420/#5423/#5424/#5425.

The package preserves the peer's witness, model, proof and accepted source gate unchanged. It adds source custody, a registered replay wrapper and explicit failure controls. The complete paid profile is {1:11077, 2:2060, 3:2550, 4:635, 5:200, 8:40}. The generous target cap is {4:140, 7:35, 8:275, 9:1590, 17:224, 25:224}; all five helper inventories are granted distinct and positive. A changed kernel, supply or degree budget needs a different analysis. This package does not duplicate the primary constructor or replay the tangent search.

New witness and source material has LOCAL provenance. Canonical selected-record and catalogue pins provide separate historical context, not fabricated ancestry for the local proposal.

Validation

The sole peer completed exactly three serial guards and four unchanged witness replays:

Actual guard	Time	Sampled peak descendant RSS	Witness calls
Registered normal and optimized	1.309556 s	73,170,944 B	2
Read-only normal	0.303061 s	35,762,176 B	1
Read-only optimized	0.504841 s	44,232,704 B	1

All guards exited zero without termination. Each replay reproduced all 3,283 raw RESULT bytes and the complete typed JSON, with zero exclusions: SHA256 4895186a36fe9c38e37da61ca752a2d07f4a57dd3160917c25e91aec334b1074. All 64 stdout bytes matched the original and stderr was empty. Stdout contains only a PASS label; the separately retained RESULT is essential. Each wrapper reported all 14 source, origin, scope, type and fraction controls.

The four wrapper pre/post digests and peer attestation record unchanged custody. The exact request/source freeze holds 42 working and 42 read-only file vectors; a separate current static readback matches them. The peer preflight contains counts and an attestation, not a newly invented expanded historical inventory. The 21 read-only directories remain mode 0555. Portable commands use an absolute interpreter, empty PATH, unset Git variables and an explicit Git discovery ceiling; ancestor Git metadata is preserved.

The exact numerical runtime includes nine files, with both B/F packets retained unchanged. Native endpoints and source applicability are not freshly reevaluated: the source gate and independent mathematical review are inherited, scoped evidence. Broader primary context is explicitly hash referenced where unbundled. The original source-gate byte verifier is not registered. No Powers, logarithm/exponential evaluation, native/phase/Lean engine, enumeration or new case ran.

The unexecuted v1 snapshot remains preserved. V2 resolved the sole pregrant gap by capturing stdout as binary and requiring its complete raw equality, alongside complete raw and typed RESULT equality.

Actual replays, independent post-run review and passive-evidence review PASS. Metadata-only validation and committed-byte/declared-source-history checks passed at both commits. The science/evidence commit contains 42 unchanged tested payloads plus two passive annexes. Its publication successor changes only docs/ci-verification.md; there are 45 PR paths in total. The checkout is clean. First hosted run 40091291250, attempt 1, passed at exact publication head a72db1fdffc9532b075f3706ed2c6e8654d6c978. Plan job 120864325122 passed onboarding and package selection; replay job 120864973219 passed the bounded normal/optimized replay and unchanged-input check. Hosted verification preserves the stated fixed-profile and inherited-source boundaries; it establishes no new κ.

The official export contains 42 payloads plus .research-export.json (43 members, 2,273,280 bytes), SHA256 8087b53f9383470ba545f8dfcf84c25c3ddaeecbb9e0f4174d644af038a9dd5b. Every payload matches the executed working/read-only copies and both commits byte for byte. Working filesystem modes were 0644, tested read-only modes 0444, and committed/exported modes 0644. The official archive was checked, not executed again. The history preflight checked the sole declared canonical pin; the available 7,507-commit metadata graph does not imply a complete historical file corpus or canonical ancestry for LOCAL sources.

Two passive public annexes retain the exact peer summary and complete raw guard/stdout/custody evidence, from which all four fresh results can be reconstructed with zero exclusions. They are outside the 42 registered inputs and the source export: certificates/research/fivefull-sixcall-package-validation-v2.json and certificates/research/fivefull-sixcall-replay-evidence-v2.json. The frozen README's preparation status is historical; these actual receipts supersede it without a source edit or scientific rerun merely to update prose.

Reproduction: python3 -B scripts/research_package.py --package research/fivefull-sixcall-complex-ceiling, Python 3.12, standard library only. That command was already run by the sole peer; this preparation lane parsed existing artifacts only. The approved metadata/history/export checks used flags whose call branches do not invoke the verifier. The reused CI-documentation repair is isolated in the publication commit; scientific source bytes are unchanged.

Credit

Matarener/Mark with Claude supplied the reported tangent lead; Cinna Davis/1337isnot1337 and PR4666 supplied the relevant theory. Credit the peer invoice/proof owner, independent source and post-run auditors, neutral5413 and native/provider/helper/phase/ordinary authors, Ali Romman, Dillon Amburgey/dillona, Cao Yifan, zxv, Dugongue, Jacob Alan Sussman, Chris R/Clear1nsight and retained compiler/BIT/analytic predecessors. This packaging uses substantial OpenAI Codex assistance. Preserve original notices and Apache-2.0/source-specific terms. Neither finite certificate success nor this scoped audit implies formal verification or complete theorem approval.

Open next questions

The required seven-label handoff points back to this PR as the canonical investigation thread.

Can a cheaper completion or a changed productive inventory evade this fixed-profile obstruction? First supply its complete changed ledger and source mapping; the actual eight-H8 bill cannot rule out alternatives. Can the BIT1920 class or a mode/coded-output caller be priced? Its own source alphabet, canonical endpoints, target independence and completion costs must be joined separately. Repeating this unchanged rational witness does not answer those distinct questions.