Plan the pre-RC patch levels

Add detailed 0.9.1 UI/dialog/button polish, 0.9.2 NPC model refinement,
0.9.3 texture/material refinement and 0.9.4 integration/performance/judge
workstreams before the 0.9.5 release candidate. Update the remaining-gap table,
baseline to v0.9.0 and README development status.

TODO.md

@@ -33,16 +33,186 @@ release-candidate foundation, not yet a 1.0 product.
 | Area | Current baseline | Remaining gap | Risk | Milestone |
 | --- | --- | --- | --- | --- |
 | Content and balance | 3 venues with validated 12+ event matrices, recovery events and contextual dialogue | External playtest evidence, balance/fairness pass, event-weight tuning and final content lock | Repetition or unfair outcomes survive into 1.0 | 0.9.5 |
-| Animation/presentation | Procedural pose layers, portrait captures, reduced hem clipping | Final animation/readability acceptance and no remaining P0/P1 clipping in the release candidate | Close-up jank reduces perceived quality | 0.9.5 |
+| Dialog/UI | Complete screens and panel art, but copy/button behavior has not had a dedicated audit | Terminology, button states, long-string layout and full screen coverage need a focused pass | Inconsistent UI undermines first-hour clarity | 0.9.1 |
-| Audio/accessibility | Procedural ambience/SFX, volume controls, captions/reduce-motion settings, searchable Help | Full accessibility/localization acceptance review, audio caption coverage and P0/P1 fixes | Public 1.0 cannot claim accessible play | 0.9.5 |
+| Animation/presentation | Procedural pose layers, portrait captures, reduced hem clipping | Final model/pose/facial readability and gear-fit acceptance before visual QA | Close-up jank reduces perceived quality | 0.9.2 |
-| Localization | English JSON table, validation and input-generated Help | Pseudo-locale/layout audit and translator workflow sign-off | Text expansion/clipping at release | 0.9.5 |
+| Audio/accessibility | Procedural ambience/SFX, volume controls, captions/reduce-motion settings, searchable Help | Full accessibility/localization acceptance review, audio caption coverage and P0/P1 fixes | Public 1.0 cannot claim accessible play | 0.9.1 / 0.9.5 |
-| Performance/platform | 24-NPC headless budget (6.9 ms), Compatibility/Web/HQ paths | Target hardware matrix, Web/HQ manual profiling, frame-time/memory evidence | Platform-specific drops or startup failures | 0.9.5 |
+| Localization | English JSON table, validation and input-generated Help | Pseudo-locale/layout audit and translator workflow sign-off | Text expansion/clipping at release | 0.9.1 / 0.9.5 |
+| Performance/platform | 24-NPC headless budget (6.9 ms), Compatibility/Web/HQ paths | Target hardware matrix, Web/HQ manual profiling, frame-time/memory evidence | Platform-specific drops or startup failures | 0.9.4 / 0.9.5 |
 | Save lifecycle | Settings/campaign/history/checkpoint files, atomic writes | Migration fixtures from 0.8/0.9, fresh-install/upgrade verification, rollback instructions | Player progress loss on upgrade | 0.9.5 |
-| QA/CI | Headless suite, smokes, screenshot compare, optional judge | Full release-candidate smoke matrix, final gallery/judge, CI/release dry run and checksums | Regressions escape the release gate | 0.9.5 |
+| QA/CI | Headless suite, smokes, screenshot compare, optional judge | Full visual judge gate, release-candidate smoke matrix, final gallery and release dry run | Regressions escape the release gate | 0.9.4 / 0.9.5 |
 | Release/legal | Unsigned desktop/macOS builds, Web/HQ packages | Signing/notarization decision, checksums, AI-asset provenance, credits, privacy/body-camera and rating review | Public release blocked by trust/policy | 0.9.5 |
-| Documentation | Architecture/ADR/status/roadmap ownership established | Final README/changelog/status/TODO/smoke-checklist sign-off and known-limitations page | Docs drift at launch | 0.9.5 |
+| Documentation | Architecture/ADR/status/roadmap ownership established | Final README/changelog/status/TODO/smoke-checklist sign-off and known-limitations page | Docs drift at launch | 0.9.4 / 0.9.5 |
 | Post-launch | Stable 0.9 feature set | Live replay, more venues, VO/music, image-to-mesh props and seasonal content | Scope creep before launch | Post-1.0 |
 
+## Patch 0.9.1 — Dialog, button and UI polish
+
+Goal: make every player-facing screen read consistently and pass the Python
+screen-design judge before any model/texture changes. No gameplay rules change.
+
+### Screen and terminology inventory
+
+- [ ] Inventory every screen/panel: start menu, venue menu, tutorial, HUD, pause,
+      settings, Help, inspection, wand, radio/emergency, wheel, shift review,
+      timeline and campaign/continue screens.
+- [ ] Record the canonical terms for admit/deny, allow/refuse, warning/blocking,
+      assistance/emergency, body camera, shift/review and campaign; remove synonyms
+      that mean the same action.
+- [ ] Audit all button labels for verb-first wording, capitalization, ordering,
+      destructive-action confirmation and consistent confirm/cancel placement.
+- [ ] List every hardcoded player-facing string and route it through
+      `content/strings.en.json`; add placeholder validation for new keys.
+- [ ] Review NPC/door-staff/radio/tutorial/help/review/campaign dialogue for
+      repetition, ambiguity, tone and localization length.
+
+### Button and panel behavior
+
+- [ ] Audit normal/hover/pressed/focus/disabled states on every button, including
+      tutorial skip, venue continue, settings apply/cancel/reset, judge/review,
+      radio send/emergency hold and Help close.
+- [ ] Verify keyboard-only and gamepad-only navigation: focus order, wrapping,
+      default focus, confirm/cancel behavior and no focus traps.
+- [ ] Verify every modal closes with Esc/right-click where documented, recaptures
+      the mouse, and does not leave the guard frozen or the menu visible.
+- [ ] Test buttons at minimum/maximum HUD font scale and with 1.5–2× longer
+      strings; fix clipping, overflow and overlapping layouts.
+- [ ] Test destructive/irreversible actions (campaign reset, tutorial replay,
+      emergency send) require the intended confirmation/hold and cannot double-fire.
+- [ ] Verify radio normal-send vs emergency hold, inspection/wand allow/deny,
+      review continue/retry/menu and Help search/filter behavior end to end.
+
+### Visual and judge verification
+
+- [ ] Update the affected UI art/reference mapping where copy or button order
+      changes; keep the no-overwrite image convention.
+- [ ] Capture a deterministic screenshot for every dialog/button state that has a
+      design reference; add missing capture stages or an explicit coverage gap.
+- [ ] Run `tools/check_screenshots.sh` and `python3 tools/screen_designs.py compare`;
+      fix all `bad` results.
+- [ ] Run `python3 tools/screen_designs.py judge` with `--model auto` (or a pinned
+      model); clear every `diff` and record accepted `review` verdicts.
+- [ ] Record the judge model, fallback order, spend and evidence in
+      `docs/screens/realism/` or the screen-design report.
+- [ ] Add a short UI copy/button checklist to `docs/smoke-checklist.md`.
+- [ ] Exit: all screens have consistent terminology and button behavior, long
+      strings/scale pass, and the Python compare/judge gate has no unexplained
+      diff.
+
+## Patch 0.9.2 — NPC model refinement
+
+Goal: make the eight character variants and their gear hold up in portraits and
+normal play without changing the `npc_base`/`Rig/Head` contract.
+
+### Model and rig audit
+
+- [ ] Render all eight variants from front, profile, back and portrait angles;
+      compare faces, hair, ears, hands, shoes, proportions and gear fit.
+- [ ] Audit head/face readability (brow, eyes, lips, nose, hairline) at portrait
+      distance; fix the highest-impact silhouette/face issues.
+- [ ] Audit gear fit and clipping: caps/beanie/hood/helmet/sunglasses, shoulder bag,
+      backpack, SCBA tank, lanyard, hand items and the new coat/dress stride.
+- [ ] Audit coat/dress/skirt hems during walk, blocking, sitting and turning; fix
+      remaining leg/hem intersections with weights, geometry or stride limits.
+- [ ] Re-check the frozen contract: mesh `npc_base`, bone names, five tint slots,
+      front direction, attach points, triangle/vertex budgets and grounding.
+- [ ] Rebuild affected GLBs, bake AO, run `godot --headless --path . --import`, and
+      run `blender_assets.py audit` plus fit/portrait checks.
+
+### Pose and animation polish
+
+- [ ] Review idle, walk, warning, blocking, recognition, refusal, phone, clipboard,
+      photo, film, sit and responder poses for pops or floating props.
+- [ ] Add/adjust short transitions for negotiation/reaction poses without breaking
+      the seeded pose helpers or head-turn behavior.
+- [ ] Verify LOD tiers and the impostor at 16/24-NPC distances; re-run the budget
+      smoke and record any triangle/draw-call changes.
+- [ ] Refresh the Blender review sheets and character portraits; keep per-asset
+      previews local if gitignored.
+
+### Verification
+
+- [ ] Add/update headless tests for any new pose/model contract behavior.
+- [ ] Capture the portrait and character gallery; compare against the design
+      references and run the Python judge.
+- [ ] Update `docs/blender-assets.md`, architecture/status and `MEMORY.md` with the
+      revised budgets, limitations and rebuild commands.
+- [ ] Exit: portrait/readability and gear fit pass, no obvious P0/P1 clipping remains,
+      and the 24-NPC budget/full test suite are green.
+
+## Patch 0.9.3 — Texture and material refinement
+
+Goal: improve close-range surface and character texture quality while keeping GL
+Compatibility/Web parity, deterministic assets and offline CI.
+
+### Environment textures and materials
+
+- [ ] Audit the v2 picks (brick, sidewalk pavers, dry/wet asphalt, awning canvas,
+      day/dusk/night skylines) at close range and in the entrance/street shots for
+      seams, tiling, scale, contrast and normal-map strength.
+- [ ] Refine brick mortar/wear, paver joints/stains, asphalt aggregate/wet response
+      and awning weave using the existing image/asset tooling; never overwrite a
+      pick, use `-vN` candidates and promote validated results.
+- [ ] Check UV scales and material roughness/metallic profiles; keep day/night
+      phase behavior and the wet-street transition intact.
+- [ ] Review distant skyline haze and foliage/planter surfaces against the concept
+      references; fix obvious flatness without regressing readability.
+
+### Character and prop textures
+
+- [ ] Audit NPC skin, cloth, pants, hair and gear albedo/normal/AO maps in portraits;
+      refine fabric weave, skin variation and fixed detail materials.
+- [ ] Re-bake AO where geometry changes; verify the 256² maps and material wiring.
+- [ ] Audit library props and UI/card/icon textures at their actual screen sizes;
+      remove blurry/oversized or unused textures where safe.
+- [ ] Run texture audits, import, and the full asset test suite; record texture memory
+      and triangle/draw-call impact.
+
+### Visual verification
+
+- [ ] Run `python3 tools/scene_concepts.py realism` for the affected stages with the
+      judge enabled; record metric gaps and model/spend.
+- [ ] Re-run `tools/check_screenshots.sh` and the screen-design judge; clear all
+      `diff` verdicts.
+- [ ] Capture before/after close-ups for brick, pavement, asphalt, character cloth
+      and skyline; document accepted trade-offs.
+- [ ] Update asset docs and the changelog/status with the promoted texture versions.
+
+## Patch 0.9.4 — Integration, performance and judge gate
+
+Goal: integrate the UI/model/texture patches, prove the combined build on the
+target paths and leave a clean release-candidate handoff.
+
+### Integration and asset rebuild
+
+- [ ] Rebuild all changed Blender/GLB/AO assets, re-import, and verify manifest/fit/
+      audit coverage.
+- [ ] Run the complete headless suite, Python tooling and smoke matrix; fix any
+      regression from the three polish patches.
+- [ ] Re-run the 24-NPC budget smoke and record frame time, mesh count and any
+      texture-memory change; check Web and Desktop HQ if available.
+- [ ] Verify save/campaign/tutorial behavior is unchanged by the visual/UI patches.
+
+### Visual gallery and judge gate
+
+- [ ] Recapture the full gallery, portraits, venue/realsim stages and Desktop HQ A/B
+      as needed; no stale screenshots in the release set.
+- [ ] Run `tools/check_screenshots.sh`, `screen_designs.py compare` and
+      `screen_designs.py judge` on the final captures.
+- [ ] Run the realism judge for the affected concepts with a pinned model; record
+      model, attempt count, spend and verdicts.
+- [ ] Review all `review` verdicts manually; convert unexplained layout/composition
+      mismatches into fixes or explicit known limitations.
+- [ ] Ensure generated realism folders remain gitignored/local-only and only the
+      durable report decisions/docs are committed.
+
+### Release-candidate handoff
+
+- [ ] Update README, architecture, development status, changelog and TODO with the
+      final patch-level results and any remaining gaps.
+- [ ] Tag/version or otherwise freeze `0.9.4` as the integration checkpoint before
+      the `0.9.5` release candidate.
+- [ ] Exit: combined build is functionally green, visually reviewed, performance
+      measured and ready for the 0.9.5 content-lock/playtest phase.
+
 ## Milestone 0.9.5 — Release candidate
 
 Goal: freeze the feature set, prove it with external players, clear release/legal blockers and