feat(M48): continuous TLS health monitoring — endpoint state machine, shared tlsprobe, 8 API endpoints, GUI

Adds continuous TLS endpoint health monitoring that closes the deploy→verify→monitor loop.
After M25 verifies a deployment succeeded once, M48 continuously confirms it stays healthy.

Key components:
- Shared `internal/tlsprobe/` package extracted from network scanner for reuse
- Health status state machine: healthy → degraded (2 failures) → down (5 failures),
  plus cert_mismatch when served fingerprint differs from expected
- 8th scheduler loop (60s tick, per-endpoint configurable intervals)
- PostgreSQL migration 000011: endpoint_health_checks + endpoint_health_history tables
- 8 REST API endpoints (CRUD, history, acknowledge, summary)
- Health Monitor GUI page with summary bar, status table, create modal, auto-refresh
- 38 new tests (5 tlsprobe + 11 domain + 10 service + 8 handler + 4 frontend)
- All coverage thresholds maintained (service 68%, handler 83%, domain 87%, middleware 63%)

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

web/src/components/StatusBadge.tsx

@@ -31,6 +31,12 @@ const statusStyles: Record<string, string> = {
   pending:             'badge-warning',
   failed:              'badge-danger',
   read:                'badge-neutral',
+  // Health check statuses
+  healthy:             'badge-success',
+  degraded:            'badge-warning',
+  down:                'badge-danger',
+  cert_mismatch:       'badge-warning',
+  unknown:             'badge-neutral',
 };
 
 export default function StatusBadge({ status }: { status: string }) {