feat/teaser-prototype-playable-core (#1)

Co-authored-by: John Lamb <j.lamb13@gmail.com>
Reviewed-on: #1

docs/brainstorms/2026-02-01-teaser-prototype-brainstorm.md

@@ -0,0 +1,67 @@
+---
+date: 2026-02-01
+topic: teaser-prototype
+---
+
+# Teaser Prototype: Playable Game Core
+
+## What We're Building
+
+A minimal playable teaser that captures the "practiced chaos" of Neighborhood Quarterback - the Rocket League-style feeling where chaos has patterns you can learn to exploit.
+
+**Scope:**
+- 1v1 vs AI opponent
+- 5 identical units per side
+- Draw-routes for commanding units
+- Tag-out respawn (captured units respawn at base after delay)
+- Simple fog of war (see near your units only)
+- Backyard-style asymmetric map with houses/fences
+- First to 3 points (points for flag grab AND returning flag to base)
+
+## Why This Approach
+
+We chose **Minimal Playable Core** over polished slice or systems-first approaches because:
+
+1. **Validate the feel first** - The soul doc's "practiced chaos" needs playtesting to verify
+2. **Fast iteration** - Rough edges are fine if we can quickly change what matters
+3. **Avoid over-engineering** - Don't build robust systems for unvalidated design
+
+## Key Decisions
+
+- **AI opponent over multiplayer**: Simpler to build, can playtest alone, control pacing
+- **Draw routes over click-to-move**: More tactical, matches the "quarterback" command fantasy
+- **Tag-out over jail escort**: Simpler first pass; jail escort adds complexity we can add later
+- **Fog of war included**: Core to the mind game, worth the complexity
+- **5 units (not 3)**: Matches soul doc, enables interesting squad tactics
+- **All identical units**: No classes yet; focus on positioning/routes before differentiation
+- **Asymmetric map**: Thematic "backyard" feel even if harder to balance
+- **First to 3 with grab+return points**: Creates multiple scoring opportunities per round
+
+## What Success Looks Like
+
+When playing, you should see:
+- Moments where you read the AI's pattern and exploit it
+- Chaotic scrambles when plans collide
+- "Almost had it" flag runs that feel learnable
+- Fog reveal moments that create tension
+
+## Out of Scope (For Now)
+
+- Class differentiation (Sneak/Patrol/Speed)
+- Jail escort mechanics
+- Motion lights
+- Pre-phase setup (placing flag/jail)
+- Multiplayer networking
+- Polish/juice/animations
+
+## Open Questions for Planning
+
+1. **Map layout**: What's the minimum topology for interesting play? Lanes, chokepoints, shortcuts?
+2. **AI behavior**: How smart does AI need to be to create "practiced chaos"?
+3. **Route-drawing UX**: Click-drag? Waypoints? How to visualize planned route?
+4. **Fog implementation**: Tile-based? Raycast? Mesh-based reveal?
+5. **Scoring flow**: What happens after a point? Reset positions? Continuous play?
+
+## Next Steps
+
+Run `/workflows:plan` to break this down into implementation tasks.