feat: Full alert engine — 17 conditions, scaling cooldown, per-condition TUI toggles

Alert conditions across all 5 pillars:
  Infrastructure: offline, recovery, new router
  Power: battery 50/25/10%, 7-day trend, USB→battery, solar not charging
  Utilization: sustained >20% for 6h, packet flood >500/24h
  Coverage: infra single gateway, feeder offline, region blackout
  Scores: mesh <70, region <60

Scaling cooldown: immediate → 12h → 24h → 48h → stop
Recovery notifications when conditions resolve
Per-condition on/off toggles in TUI
Battery trend queries SQLite node_snapshots for 7-day history

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

meshai/main.py

@@ -279,10 +279,10 @@ class MeshAI:
                 health_engine=self.health_engine,
                 reporter=self.mesh_reporter,
                 subscription_manager=self.subscription_manager,
-                critical_nodes=getattr(mi, 'critical_nodes', []),
-                alert_cooldown_minutes=getattr(mi, 'alert_cooldown_minutes', 30),
+                config=mi,
+                db_path="/data/mesh_history.db",
             )
-            logger.info(f"Alert engine initialized (critical nodes: {getattr(mi, 'critical_nodes', [])})")
+            logger.info(f"Alert engine initialized (critical: {mi.critical_nodes}, channel: {mi.alert_channel})")
 
         # Knowledge base (optional - gracefully degrade if deps missing)
         kb_cfg = self.config.knowledge