fix: agent timeouts hid hf-pdocker-01; refresh links and groups

- agent: stats one_shot=True + 20 workers (.192 went 12s -> ~2s);
  /health ignores Dead containers
- server: /containers fetch timeout 10s -> 30s (was silently returning
  [] for the busiest node)
- links page: drop FRIDAY Voice, add Onyx, Guacamole, Synapse Admin,
  n8n, CouchDB; MCPO -> /docs, Trivy -> /healthz, Jellyseerr -> Seerr
- config.json: drop dead gpu-workstation (192.168.86.27)
- groups.json: rebuilt for current container names (14 groups)
- includes earlier uncommitted work (models, styles, ports sorting)

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01NQhgEUqcaV4WM2S4hfXgHD

server/static/style.css

@@ -561,6 +561,25 @@ body {
     text-overflow: ellipsis;
 }
 
+.card-ports {
+    display: flex;
+    flex-wrap: wrap;
+    gap: 4px;
+    overflow: visible;
+    white-space: normal;
+}
+
+.port-badge {
+    display: inline-block;
+    background: var(--bg-tertiary);
+    border: 1px solid var(--border);
+    color: var(--accent);
+    font-size: 11px;
+    font-family: monospace;
+    padding: 1px 6px;
+    border-radius: 4px;
+}
+
 .card-actions {
     display: flex;
     gap: 6px;