Ideation: large-scale fixes and expansions

Restored 2026-08-03 from the rescue/uncommitted-2026-07-05 snapshot; a historical planning record. Terminal states for every item named here live in docs/ROADMAP-CLOSEOUT.md, and active planning lives in docs/NOVEL-USE-CASES-PLAN.md and docs/ROADMAP-MULTIYEAR.md.

Drafted 2026-07-01. This folder is the third documentation layer for constituent-reconciler, and the most speculative one. The layers relate this way:

• ../ROADMAP.md is the canonical build spec: the shipped phases (v0.1 through v0.7), the architecture, the eval plan, the metrics ledger, and the 1.0 gate.
• ../RESEARCH-ROADMAP.md (2026-06-30) is the persona-derived backlog: remediation items R1 to R11 and expansions E1 to E10, triaged from the synthetic panel in ../USER-RESEARCH.md.
• This folder is a deep-dive ideation pass over the actual code as it stands at v0.7 (commit 88dc26a). It proposes structural fixes and expansions that the two documents above do not already contain. Where an idea builds on an existing item it cites that item by ID (R1 to R11, E1 to E10) and states what is new; nothing here restates an item those files already own.

Contents

File	What it holds
01-deep-dive.md	Current-state assessment from a fresh read of the source: architecture, genuine strengths, observed structural debt, and the repo's position in the portfolio.
02-large-scale-fixes.md	FIX-01 to FIX-12: deep structural fixes across correctness, security, privacy, operability, and maintainability, each with an effort tier and a measurable bar.
03-expansions.md	EXP-01 to EXP-16 in three horizons: deepen the core, adjacent capabilities, transformative bets.
04-impact-and-sequencing.md	Impact-by-effort matrix over all IDs, dependency notes, a Now/Next/Later sequence that extends the existing roadmaps, and the items gated on human, legal, SME, or real-data input.

What this folder is not

These are ideas for evaluation, not commitments. Nothing here has been scoped with users, sized against maintainer capacity, or reviewed by counsel. Several items carry explicit gates (counsel review, a real-organization pilot, an accessibility SME) and the honest position of this project is that gated work is deferred and reported as deferred, never simulated. Priorities in 04-impact-and-sequencing.md are the author's read of the code, made on one day, and should be treated the way RESEARCH-ROADMAP.md treats its own priorities: as hypotheses.