feat(dashboard): add HTTP health probe for cross-container gateway detection

The dashboard's gateway status detection relied solely on local PID checks
(os.kill + /proc), which fails when the gateway runs in a separate container.

Changes:
- web_server.py: Add _probe_gateway_health() that queries the gateway's HTTP
  /health/detailed endpoint when the local PID check fails. Activated by
  setting the GATEWAY_HEALTH_URL env var (e.g. http://gateway:8642/health).
  Falls back to standard PID check when the env var is not set.
- api_server.py: Add GET /health/detailed endpoint that returns full gateway
  state (platforms, gateway_state, active_agents, pid, etc.) without auth.
  The existing GET /health remains unchanged for backwards compatibility.
- StatusPage.tsx: Handle the case where gateway_pid is null but the gateway
  is running remotely, displaying 'Running (remote)' instead of 'PID null'.

Environment variables:
- GATEWAY_HEALTH_URL: URL of the gateway health endpoint (e.g.
  http://gateway-container:8642/health). Unset = local PID check only.
- GATEWAY_HEALTH_TIMEOUT: Probe timeout in seconds (default: 3).

web/src/pages/StatusPage.tsx

@@ -53,7 +53,8 @@ export default function StatusPage() {
   };
 
   function gatewayValue(): string {
-    if (status!.gateway_running) return `${t.status.pid} ${status!.gateway_pid}`;
+    if (status!.gateway_running && status!.gateway_pid) return `${t.status.pid} ${status!.gateway_pid}`;
+    if (status!.gateway_running) return t.status.runningRemote;
     if (status!.gateway_state === "startup_failed") return t.status.startFailed;
     return t.status.notRunning;
   }