# Reconciliation: Brainstorming → PRD

## Gaps Found (5 total)
### G-1: Reaction Cam consent model flipped from opt-in to opt-out
**Source:** Brainstorming Phase 3 / Sofia persona + Key Tension Resolved (brainstorming-session-2026-05-19-221747.md:130-150)
**Gap:** The brainstorming resolved the Marcus/Sofia tension with a consent-aware rule: Reaction Cam is **opt-in at setup**. The PRD partially preserves the privacy concern, but FR-24 defines Reaction Cam as an **opt-out** setting, which weakens the explicit consent model and contradicts the brainstorming resolution.
**Severity:** Critical
**Suggestion:** Update the PRD so Reaction Cam is consistently framed as opt-in everywhere (Vision, Privacy Panel, Roadmap, and any future automation language). If the product owner intentionally changed this, document the rationale explicitly as a scope decision rather than leaving it as a silent inversion.

### G-2: Vocabulary-tier UX and product voice are under-specified
**Source:** Alex persona + Idea #21 / Week 1–2 priorities (brainstorming-session-2026-05-19-221747.md:142-146, 173-178, 225-230)
**Gap:** The brainstorming made "plain language by default, power terms in advanced mode" a distinct product idea and onboarding strategy. The PRD keeps some plain-language copy examples, but it does not preserve the broader UX rule of **default-simple vocabulary with optional advanced/cinematic terminology**. Because the PRD is glossary-heavy, downstream implementation could drift toward expert-facing labels and lose the beginner-friendly voice that Alex validated.
**Severity:** Moderate
**Suggestion:** Add an explicit UX/content principle or requirement: default labels use plain language (Show/Hide/Spotlight), while cinematic or advanced terminology is optional in advanced mode/tooltips/documentation.

### G-3: Director's Board usage model drifted away from the brainstormed workflow
**Source:** SCAMPER Combine decision + Marcus/Jake personas (brainstorming-session-2026-05-19-221747.md:101, 124-140)
**Gap:** The brainstorming positioned the seating-chart popout as a **low-frequency setup/control tool**: open at session start, close during play, and reopen instantly via shortcut when needed. The PRD includes keyboard access, but UJ-3 reframes the Director's Board as something that stays open in a corner during the session. That changes the intended operating model and may push the implementation toward a noisier, always-on control-room UI instead of a lightweight prep-time tool.
**Severity:** Moderate
**Suggestion:** Revise the PRD to clarify the intended workflow: the Director's Board is primarily a pre-session / between-scenes tool with instant shortcut recall, not something assumed to remain open throughout play. If both modes are desirable, state that explicitly.

### G-4: "Every automation needs a one-click GM override" is not elevated as a cross-cutting rule
**Source:** Marcus persona + Question Storming "panic button" (brainstorming-session-2026-05-19-221747.md:54, 124-128)
**Gap:** The brainstorming established a strong control principle: automation is welcome only if the GM can immediately override it, including an instant hide-all/panic path. The PRD implements some related mechanics (Hide All, manual overrides for Scene Presets), but it does not carry forward the broader **product rule** that every present and future automation surface must expose an immediate GM override. That omission matters because downstream work on Scene Presets, Combat Cinematics, and Reaction Cam could become harder to interrupt safely.
**Severity:** Moderate
**Suggestion:** Add a cross-cutting requirement or design principle: every automation-capable feature must provide an obvious one-click GM override, and the module must expose an emergency hide-all action reachable without opening deep configuration UI.

### G-5: The brainstorm's phased-delivery discipline is blurred in the PRD
**Source:** North Star + Priority Stack + Session Summary (brainstorming-session-2026-05-19-221747.md:208-238, 277-293)
**Gap:** The brainstorming repeatedly anchored the product around a tightly scoped Day 1 ship: right-click toggle first, then practical Week 1–2 enhancements, then cinematic/streaming power features later. The PRD preserves the concepts but largely collapses Day 1 and Week 1–2 work into one v1.0 scope. That weakens the brainstorming's explicit risk-control signal: keep the first release rooted in the shippable core toggle.
**Severity:** Moderate
**Suggestion:** Add a release-phasing addendum or split scope more clearly into MVP / post-MVP milestones (for example: Level 1 ship first, then Director's Board + Presets, then privacy/automation scaffolding). If the all-in v1.0 scope is intentional, note the increased delivery risk explicitly.

## Intentional Drops (confirmed out of scope)
- Full asymmetric per-viewer visibility control / who-sees-who UI exposure — correctly deferred to Later roadmap / v1.0 non-goal.
- Pull Visibility Model — correctly deferred to Later roadmap.
- Combat Cinematics Mode — correctly deferred to Later roadmap.
- Reaction Cam feature execution — correctly deferred to Later roadmap (while privacy prerequisites remain relevant now).
- Stage Lighting States / theatrical vocabulary mode — correctly deferred to Later roadmap.
- Token-Anchored Floating Cams — correctly deferred to Later roadmap.
- HP-Reactive Camera Styling runtime behavior — correctly deferred to Later roadmap.
- NPC Presence Tiles — correctly deferred to Later roadmap.
- Spectator Curtain / Dual Layout System — correctly deferred to Later roadmap.
- Browser Source API / OBS-ready tile URLs — correctly deferred to Later roadmap.
- Zero-UI Full Automation Mode — correctly deferred to Later roadmap.
- Reaction Clip System — correctly deferred to Later roadmap.
- Non-native AV backends (e.g. Jitsi) — correctly excluded for v1.0 by explicit non-goal/open-question resolution.