feat: Implement adaptive leverage based on signal quality score

- Quality-based risk adjustment: 95+ = 15x, 90-94 = 10x, <90 = blocked
- Data-driven decision: v8 quality 95+ = 100% WR (4/4 wins)
- Config fields: useAdaptiveLeverage, highQualityLeverage, lowQualityLeverage, qualityLeverageThreshold
- Helper function: getLeverageForQualityScore() returns appropriate leverage tier
- Position sizing: Modified getActualPositionSizeForSymbol() to accept optional qualityScore param
- Execute endpoint: Calculate quality score early (before sizing) for leverage determination
- Test endpoint: Uses quality 100 for maximum leverage on manual test trades
- ENV variables: USE_ADAPTIVE_LEVERAGE, HIGH_QUALITY_LEVERAGE, LOW_QUALITY_LEVERAGE, QUALITY_LEVERAGE_THRESHOLD
- Impact: 33% less exposure on borderline quality signals (90-94)
- Example: $540 × 10x = $5,400 vs $8,100 (saves $2,700 exposure on volatile signals)
- Files changed:
  * config/trading.ts (interface, config, ENV, helper function, position sizing)
  * app/api/trading/execute/route.ts (early quality calculation, pass to sizing)
  * app/api/trading/test/route.ts (quality 100 for test trades)

app/api/trading/test/route.ts

@@ -74,11 +74,13 @@ export async function POST(request: NextRequest): Promise<NextResponse<TestTrade
     }
     
     // Get symbol-specific position sizing (with percentage support)
+    // Test trades use quality score 100 for maximum leverage (manual override)
     const { getActualPositionSizeForSymbol } = await import('@/config/trading')
     const { size: positionSize, leverage, enabled, usePercentage } = await getActualPositionSizeForSymbol(
       driftSymbol, 
       config, 
-      health.freeCollateral
+      health.freeCollateral,
+      100 // Test trades always use max leverage (quality 100)
     )
     
     // Check if trading is enabled for this symbol