Research reset: expose geometric consistency, source-visible state, and the original-U bottleneck

[LightChainr]

Owner-delegated research direction reset — 2026-09-13

This is the navigation/allocation half of the owner's request to actively reshape the repository. The actual new mathematical result is submitted separately as draft #736; it is not smuggled into this merge.

What changes

• Rewrite README.md, docs/ROADMAP.md, docs/RESEARCH-MAP.md and docs/STATUS.md around three distinct research questions, with corrected current scope instead of stale acquisition instructions.
• Preserve the exact four previous Git blobs under docs/history/*-before-20260913.md. The archive index points to the original tree for historical relative links. No original raw data, result, scorer, frozen design or research branch is removed or overwritten.
• Add docs/RESEARCH-FRONTIER.md: pinned unmerged research assets, completed-versus-open distinctions, a correction map, and the new axial full-law theorem's role.
• Add lightweight AGENTS.md and analysis/research_agenda.json. The JSON is allocation-only, not a second evidence ledger, scheduler or enforcement framework.

Research allocation

Primary #613: geometric separation of root and full-law consistency. #736 has already delivered axial necessity; the next question is the uniform oblique winding-corridor lemma, with independent proof examination and targeted novelty comparison rather than more census.

Bounded parallel #636: actual site-transfer closure weights and source-visible cancellations beyond the completed #708/#710/#733 width-four laboratory. No automatic new width ladder.

Theory-blocked flagship #275, P1: retain the original source/normalizer/thermal-jet/moving-root contract. Obtain two actual forward maps and check represented, nuisance-profiled prediction images before sampling. Changing its resource priority is not rejecting the physical question.

#537's remaining proof obligations and the publication units remain visible. #276/#575/#583 are not reopened. #622's #706 delivery is not commissioned again. Live issue descriptions/labels are updated separately under the same delegation.

Scope and verification

12 text/JSON paths, including the four exact archive copies. Existing executable source, test, workflow, dependency and numerical-result paths are unchanged. New JSON parsed locally; text and cited asset refs checked. No document-wording tests or audit machinery added. Full repository CI was not run for this navigation delta; no CI success is claimed. #736's three local mathematical tests and exact controls belong to that separate research PR, not this one.

This reset does not certify all open research PRs, claim original-U identification, or assert publication novelty. The existing unmerged stack remains unmerged. Main integration of this navigation change is requested under the owner's explicit instruction to take control of the research direction.

[LightChainr]

Post-merge allocation refinement: after reading the newer #650 single-active-deliverable decision, main was fast-forwarded (no force) to d31fa5fe9584b77595e4a78b9560aa2ff77cf0b3 with documentation/JSON changes only. This supersedes this PR body's preliminary bounded parallel allocation: one probability acceptance package is active, consolidating/auditing #735 and incorporating #613/#736's sharp axial full-law boundary. #636 is a visible P1 representation reserve, paused for new extension; #275 is a P1 theory-blocked flagship with its original contract unchanged. The optional oblique-corridor extension is not a new prerequisite or second dispatch. Current main's FRONTIER/ROADMAP/AGENTS/agenda and updated #613/#636 agree. The newer #650 body was preserved and an execution handoff was added there (comment 5647465417), rather than overwritten. Concurrent #737 was read and incorporated with its exact finite scope and disclosed delivery-input gaps. No research stack, including new #736, was merged.