EPIC: USER-TRUTH GATES — the paint the player sees must agree with the game they play (owner playtest 2026-07-22)

[100yenadmin]

Owner playtest verdict (live player, screenshots on file): walk through walls/staircases/objects/fire, invisible character, spawn inside a bookcase, no functional doors — 'it doesn't make sense from a user perspective.' Engine probes on the same instance show the ENGINE layer correct (walls/props reject imp=Y, camera contract exact). The gap is systematic: every gate we run certifies GEOMETRY truth; NOTHING certifies the USER'S VISUAL truth. The five feedback items decompose into 4 instrument gaps + 1 stale build (fixed):

1. FIXED TONIGHT — invisible/crumpled actor: the client: runtime-spawned actors bind the animated roster setup (T-pose fix, #1601) #1616 fallback-model fix existed in code but no build carrying it had reached the owner app. Installed (with a new refresh-flow trap: ditto-to-~/Applications → OS_REASON_CODESIGNING kill; fix = xattr -dr quarantine + ad-hoc codesign — runbook update queued). Post-install frame: actor upright + animated.
2. Occluded-actor silhouette (P1): standing behind a tall occluder fully masks the sprite — only the name label floats ('invisible half the time'). Isometric convention: draw a tinted SILHOUETTE through occluders. Client change; rides the next box build.
3. Door affordance (P1 — escalates Room Pipeline: door-cell ↔ painted-arch reconciliation (cosmetic) #1585 from 'cosmetic'): the door CELL works but the PAINTED archway is elsewhere; clicking the visible door does nothing → 'no functional doors.' Fix: click-target at the painted arch mapped to the door cell (per-room arch-pixel annotation or glow hotspot at the arch), not just the label.
4. Paint-coverage coherence (P1): grid-open cells under painted furniture ('spawn in a bookcase', 'walk through objects/fire', 'through the staircase' — that staircase is paint with no geometry). Reproduced live: engine happily walks to (1,1) in the shop = inside the painted shelf. Fix: extend check_grid_paint_coherence (the camp-truegrey instrument) into a PER-ROOM CERTIFICATION STAGE — every walkable cell must be visually floor-open in the plate (or the cell gets blocked / the plate re-locked); spawns + door arrivals REQUIRE paint-open cells.
5. USER-TRUTH walk-gate stage: the walk gate gains: click-the-painted-door (must cross), walk-behind-occluder (silhouette visible), spawn/arrival visual-openness. A room certifies only when the GAME THE USER SEES agrees with the game the engine plays.
6. Owner-instance continuous gate: gates ran in the QA sandbox on new builds while the owner pair went unverified between installs. Rule: any owner-touching change ends with a walk-gate probe against the OWNER instance (probe-set + party-return), and installs end with a fresh-frame eyeball.

Companions: #1639 (cast presence), #1645 (combat lifecycle), #1644 (camp plate retirement). This epic is the next sprint's spine — it outranks further paint cycles.

[evaos-code-review-bot]

evaOS issue enrichment

Issue: #1647 - EPIC: USER-TRUTH GATES — the paint the player sees must agree with the game they play (owner playtest 2026-07-22)
State: open

Related issues/PRs: #1616, #1585, #1639, #1645, #1644.
Existing labels: bug.
Suggested labels: docs, tests.
Suggested owners: none.

Related context

• client: runtime-spawned actors bind the animated roster setup (T-pose fix, #1601) #1616 - mentioned in issue metadata; inspect for dependency, duplicate, or prior-decision risk.
• Room Pipeline: door-cell ↔ painted-arch reconciliation (cosmetic) #1585 - mentioned in issue metadata; inspect for dependency, duplicate, or prior-decision risk.
• Walked world renders empty of its cast: rest-surface tokens exclude NPCs + spawned monsters (first A-G weakest-link finding) #1639 - mentioned in issue metadata; inspect for dependency, duplicate, or prior-decision risk.
• Text-arc weakest link: DM narrates combat without closing the engine lifecycle (no end_combat / XP / action economy) — first live arc stalls in-dungeon #1645 - mentioned in issue metadata; inspect for dependency, duplicate, or prior-decision risk.
• Retire paint-first plates (camp_clearing / camp_clearing_night): invisible-collision class — regenerate camp through the modern registration chain #1644 - mentioned in issue metadata; inspect for dependency, duplicate, or prior-decision risk.

Agent-start packet

Problem shape: EPIC: USER-TRUTH GATES — the paint the player sees must agree with the game they play (owner playtest 2026-07-22) - **Owner playtest verdict (live player, screenshots on file): walk through walls/staircases/objects/fire, invisible character, spawn inside a bookcase, no functional doors — 'it doesn't make sense from a user
Product fit: Product/UX-sensitive issue; review against repo vision and user flow quality, not only technical P0/P1 risk.
Implementation wedge: Ship the smallest user-visible path or prototype that can be smoke-tested against the intended experience.

Build / borrow / buy scan:

• Build: identify the smallest implementation wedge that proves the issue without broad rollout.
• Borrow: inspect same-repo prior art, repo memory, GitNexus context, and allowlisted cross-repo examples first.
• Buy/use: run capped external OSS/library/API/current-market research with citations, observed dates, and why each source matters.

Candidate sources:

• VISION.md / repo policy / repo-memory.md / GitNexus packet.
• Same-repo issues and PRs referenced by issue metadata.
• Allowlisted cross-repo GitHub examples with direct issue/PR/repo URLs.
• External OSS repositories or library/API docs with freshness dates.
• Current market/product examples, capped and cited, only when they change implementation direction.

Acceptance criteria:

• Add explicit acceptance criteria before implementation is considered done.
• Include source citations for any external OSS/library/current-market claims.
• Keep issue-enrichment comments sticky and idempotent.
• Keep labels, owners, reviewers, and roadmap fields suggestion-only.

Proof plan:

• Focused unit/fixture test for the changed planner behavior.
• Dry-run issue-enrichment evidence packet before live comment posting.
• Redaction check for issue body, citations, and generated markdown.
• UX/product smoke or screenshot proof when implementation changes user flow.

Known traps:

• Do not treat related links as proof by themselves; each source needs a reason why it matters.
• Do not scan old issue backlogs or exceed repo-level enrichment throttles.
• External research must stay capped, cited, fresh-dated, and redacted.
• Technical severity alone can miss product/UX regressions; include product-manager review framing.

Non-goals:

• Do not auto-apply labels, owners, reviewers, roadmap fields, or milestones.
• Do not bulk-enrich old backlog issues from this planner output.
• Do not claim external market or OSS research was performed unless cited sources are present.

Context-source taxonomy:

• [enabled] vision_repo_policy_memory_gitnexus: Use repo vision, policy, memory, and GitNexus first before external research.
• [enabled] same_repo_issues_prs: Issue metadata contains same-repo references.
• [enabled] allowlisted_cross_repo_github: Triggered by issue class; keep to configured allowlist and source caps.
• [enabled] external_oss_examples: Use for build/borrow/buy precedent with citations and observed dates.
• [enabled] library_api_docs: Use when implementation may depend on a module, SDK, API, or framework choice.
• [enabled] current_market_examples: Use for product/UX/market positioning issues with freshness dates.
• [enabled] internal_mcp_docs: Use internal docs/MCP context when available; missing context is degraded, not blocking.

Validation suggestions

• Confirm owner, acceptance criteria, and validation evidence before implementation.

Triage gaps

• Acceptance criteria or validation evidence not detected in issue body.

No labels, owners, reviewers, or roadmap fields were changed by this bot.

[100yenadmin]

Wave-2 design note (from #1652's arrivals finding): door-ARRIVAL cells are picked engine-side at cross_door time (nearest reachable free cell to the reciprocal door) — the engine must stay paint-blind, so arrival coherence lands as SEED-TIME arrival HINTS (coherence-open cells baked into world data) rather than engine paint-awareness. Also wave-2: pass the coherence-reports dir into paint_coherence.derive_room so the gate reproduces post-fix spawns.

[100yenadmin]

§9 DEMO GATE G2: MET (2026-07-22). N=3 arc-duo aggregate adv_agg_n3_20260722 (scores_db, surface=adventure): completion 1.00 · behavioral 1.00 · engagement 0.83 · story 0.76 · mechanics 0.71. All three runs completed the full Diablo-1 arc (beats 14/15/13), zero beat_failed rows, zero claude-mem refs across every DM transcript (the #1656→#1657→#1659→#1661 hermetic chain held under repeated rc=124 combat-beat timeouts — the retry machinery recovered every one). G3 was met earlier today (walk4 6/6 with full cast). Remaining: G1 (player_cert live half, #1651) + G4 (owner playthrough; polish punch list #1665/#1666). Known instrument note: the aggregate's 'stuck 0.00' weakest-link line is the #1670 entered_dungeon stamp miscalibration — the true lowest dimensions are pace (0.07: completions consume nearly the full beat budget; DM combat turns frequently brush the per-beat timeout) and mechanics (0.71).