Fix: replace Gemini vision references with MiniMax latest

pi-flow-native-brain.html: gemini-2.5-flash → minimax-latest for vision judging, GEMINI_API_KEY → MINIMAX_API_KEY, updated 'DeepSeek can't do multimodal' to reflect DeepSeek V4 now has multimodal but MiniMax is used for vision.

pi-flow-native-brain.html

@@ -457,7 +457,7 @@
   </div>
   <div class="kitchen-col">
     <span class="kitchen-col__title kitchen-col__title--reality">In the Flow</span>
-    <p><span class="gate gate--visual">G5 · Visual</span> Captures 8 frames at 100ms intervals, grids them, feeds to <code>gemini-2.5-flash</code>. Checks: T-pose? Foot-slide? Frozen animation? Wrong clip? Missing transitions?</p>
+    <p><span class="gate gate--visual">G5 · Visual</span> Captures 8 frames at 100ms intervals, grids them, feeds to <code>minimax-latest</code>. Checks: T-pose? Foot-slide? Frozen animation? Wrong clip? Missing transitions?</p>
   </div>
 </div>
 
@@ -570,7 +570,7 @@ steps:
   # G0: Pre-flight — validate vision CAN run before any build work
   - id: preflight
     agent: vision-preflight
-    task: Check GEMINI_API_KEY is set AND game_frames reaches a live instance.
+    task: Check MINIMAX_API_KEY is set AND game_frames reaches a live instance.
           If EITHER fails, STOP — vision is not optional.
 
   # Context + plan
@@ -596,7 +596,7 @@ steps:
   - id: tests       → agent: test-runner
   - id: behavior    → agent: behavioral-prober (drives LIVE game via drive_game)
   - id: feel        → agent: feel-judge (apex, airtime, latency, rise/fall)
-  - id: visual      → agent: animation-vision-judge (multimodal gemini-2.5-flash)
+  - id: visual      → agent: animation-vision-judge (multimodal minimax-latest)
 
   # Self-recurring fix-loop: bounded loop back to implement with evidence
   - id: fix-loop
@@ -612,13 +612,13 @@ steps:
 
 <p>Eighteen steps, seven cooks, five inspection points, one head chef. Triggered by a single order ticket.</p>
 
-<p>Here's how the brigade actually worked. The <strong>vision-preflight</strong> agent — the chef who checks the gas is on before anyone starts cooking — verified <code>GEMINI_API_KEY</code> was set and <code>game_frames</code> could reach the live game. Both green in under a second. Without this, the whole kitchen would prep for an hour only to discover the oven doesn't work.</p>
+<p>Here's how the brigade actually worked. The <strong>vision-preflight</strong> agent — the chef who checks the gas is on before anyone starts cooking — verified <code>MINIMAX_API_KEY</code> was set and <code>game_frames</code> could reach the live game. Both green in under a second. Without this, the whole kitchen would prep for an hour only to discover the oven doesn't work.</p>
 
 <p>The <strong>project-context-reader</strong> — the commis who reads the entire recipe book — ingested <code>PlayerController.cs</code>, <code>PlayerAnimController.cs</code>, <code>PlayerAnimationLogic.cs</code>, the test files, the manifest. The <strong>feature-planner</strong> — the sous-chef who breaks down the order into station tasks — decomposed 19 failures into four fix groups: vegetation manifest (146 broken <code>prefabPath</code> items), animation controller (crouch parameter not plumbed), jump physics (coyote time, variable height, air control — all missing), and animation tree (entire state machine absent).</p>
 
 <p>Then the <strong>game-builder</strong> — the line cook at the hot station — read each test failure like a dish ticket, traced it to the source, and started cooking. Coyote time: 100ms grace period after feet leave the ground. Variable jump height: velocity scaled by hold duration, tap gives 3.5, full hold gives 6.5. Air control: horizontal speed cut 40% while airborne. Jump phases: minimum 0.15s on jump_start before transitioning up. Landing timer: wait the full animation length, not length-minus-blend. Animation tree: <code>jump_start → jump → jump_land</code> states with 0.1s blends.</p>
 
-<p>Then the inspection line: <strong>build-verifier</strong> compiled. <strong>Test-runner</strong> ran the suite. <strong>Behavioral-prober</strong> sent <code>{"jump":true}</code> to the live game and sampled the player body. <strong>Feel-judge</strong> measured apex, airtime, liftoff latency. <strong>Animation-vision-judge</strong> captured 8 frames, gridded them, had <code>gemini-2.5-flash</code> scan for T-poses and foot-slide.</p>
+<p>Then the inspection line: <strong>build-verifier</strong> compiled. <strong>Test-runner</strong> ran the suite. <strong>Behavioral-prober</strong> sent <code>{"jump":true}</code> to the live game and sampled the player body. <strong>Feel-judge</strong> measured apex, airtime, liftoff latency. <strong>Animation-vision-judge</strong> captured 8 frames, gridded them, had <code>minimax-latest</code> scan for T-poses and foot-slide.</p>
 
 <p>Anything red → ticket back to the cook with the specific failure → fix → re-enter the line. Bounded to 3 returns. Anything green → falls through. All green → <strong>game-judge</strong> gives the final verdict.</p>
 
@@ -751,14 +751,14 @@ Context: ${{input.context}}</code></pre>
     <tr style="border-bottom:1px solid #1c2230;"><td style="padding:7px 12px;color:#e6edf3;"><code>@planning</code></td><td style="padding:7px 12px;color:#f59e0b;">DeepSeek V4</td><td style="padding:7px 12px;color:#cdd7e2;"><strong>Boning knife</strong> — precision decomposition. Breaks tasks into steps, designs DAGs. Flow architect, feature planner.</td></tr>
     <tr style="border-bottom:1px solid #1c2230;"><td style="padding:7px 12px;color:#e6edf3;"><code>@fast</code></td><td style="padding:7px 12px;color:#38bdf8;">DeepSeek V4 Flash</td><td style="padding:7px 12px;color:#cdd7e2;"><strong>Paring knife</strong> — quick, decisive cuts. Gate pass/fail, fork choices, loop exits. No overthinking.</td></tr>
     <tr style="border-bottom:1px solid #1c2230;"><td style="padding:7px 12px;color:#e6edf3;"><code>@research</code></td><td style="padding:7px 12px;color:#f59e0b;">DeepSeek V4</td><td style="padding:7px 12px;color:#cdd7e2;"><strong>Fillet knife</strong> — flexible, follows contours. Reads codebase, traces patterns, finds what matters.</td></tr>
-    <tr style="border-bottom:1px solid #1c2230;"><td style="padding:7px 12px;color:#e6edf3;"><code>@vision</code></td><td style="padding:7px 12px;color:#a855f7;">Gemini 2.5 Flash</td><td style="padding:7px 12px;color:#cdd7e2;"><strong>The inspector's eyes</strong> — the only knife that sees. Multimodal frame judging: T-poses, foot-slide, frozen anims.</td></tr>
+    <tr style="border-bottom:1px solid #1c2230;"><td style="padding:7px 12px;color:#e6edf3;"><code>@vision</code></td><td style="padding:7px 12px;color:#a855f7;">MiniMax latest</td><td style="padding:7px 12px;color:#cdd7e2;"><strong>The inspector's eyes</strong> — the only knife that sees. Multimodal frame judging: T-poses, foot-slide, frozen anims.</td></tr>
     <tr><td style="padding:7px 12px;color:#e6edf3;"><code>@compact</code></td><td style="padding:7px 12px;color:#38bdf8;">DeepSeek V4 Flash</td><td style="padding:7px 12px;color:#cdd7e2;"><strong>Kitchen shears</strong> — lightweight, versatile. Summaries, verdicts, post-processing. Fast and cheap.</td></tr>
   </tbody>
 </table>
 
 <div class="callout callout--amber">
   <span class="callout__kicker">Why DeepSeek?</span>
-  <p>Two reasons. <strong>It's free</strong> — no usage limits, which matters when your game-builder reads 800-line files and writes 200-line diffs ten times a session. <strong>It's genuinely good at C# and Godot</strong> — I've had it write a full lighting module for our Godot fork by reading Unity API docs and adapting patterns. No agent had pulled that off before. DeepSeek can't do multimodal, so vision goes to Gemini — but for everything else, it's the chef's knife you reach for 90% of the time.</p>
+  <p>Two reasons. <strong>It's free</strong> — no usage limits, which matters when your game-builder reads 800-line files and writes 200-line diffs ten times a session. <strong>It's genuinely good at C# and Godot</strong> — I've had it write a full lighting module for our Godot fork by reading Unity API docs and adapting patterns. No agent had pulled that off before. DeepSeek V4 now has multimodal, but for vision we use MiniMax latest — it's sharper at frame-by-frame animation judging and costs less per image. For everything else, DeepSeek is the chef's knife you reach for 90% of the time.</p>
 </div>
 
 <p>The point of the knife rack: you configure this <strong>once</strong>. Every agent declares <code>model: @coding</code> and gets DeepSeek V4 automatically. Swap models globally without touching any flow or agent file. The right blade, every time, no thinking required.</p>