feat(scheduler): plugin architecture + Doctor health check plugin

- Create ISchedulerJobPlugin abstraction with JobKind routing
- Add SchedulerPluginRegistry for plugin discovery and resolution
- Wrap existing scan logic as ScanJobPlugin (zero behavioral change)
- Extend Schedule model with JobKind (default "scan") and PluginConfig (jsonb)
- Add SQL migrations 007 (job_kind/plugin_config) and 008 (doctor_trends table)
- Implement DoctorJobPlugin replacing standalone doctor-scheduler service
- Add PostgresDoctorTrendRepository for persistent trend storage
- Register Doctor trend endpoints at /api/v1/scheduler/doctor/trends/*
- Seed 3 default Doctor schedules (daily full, hourly quick, weekly compliance)
- Comment out doctor-scheduler container in compose and services-matrix
- Update Doctor architecture docs and AGENTS.md with scheduling migration info

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

devops/compose/docker-compose.stella-ops.yml

@@ -1606,29 +1606,32 @@ services:
       <<: *healthcheck-tcp
     labels: *release-labels
 
-  doctor-scheduler:
-    <<: *resources-light
-    image: stellaops/doctor-scheduler:dev
-    container_name: stellaops-doctor-scheduler
-    restart: unless-stopped
-    depends_on: *depends-infra
-    environment:
-      ASPNETCORE_URLS: "http://+:80"
-      <<: [*kestrel-cert, *router-microservice-defaults, *gc-light]
-      ConnectionStrings__Default: *postgres-connection
-      ConnectionStrings__Redis: "cache.stella-ops.local:6379"
-      Router__Enabled: "${DOCTOR_SCHEDULER_ROUTER_ENABLED:-true}"
-      Router__Messaging__ConsumerGroup: "doctor-scheduler"
-    volumes:
-      - *cert-volume
-    healthcheck:
-      test: ["CMD-SHELL", "bash -c 'echo > /dev/tcp/$(hostname)/80'"]
-      <<: *healthcheck-tcp
-    networks:
-      stellaops:
-        aliases:
-          - doctor-scheduler.stella-ops.local
-    labels: *release-labels
+  # DEPRECATED: doctor-scheduler is replaced by the DoctorJobPlugin in the Scheduler service.
+  # Scheduling and trend storage are now handled by the Scheduler with jobKind="doctor".
+  # See: SPRINT_20260408_003_JobEngine_scheduler_plugin_architecture.md
+  # doctor-scheduler:
+  #   <<: *resources-light
+  #   image: stellaops/doctor-scheduler:dev
+  #   container_name: stellaops-doctor-scheduler
+  #   restart: unless-stopped
+  #   depends_on: *depends-infra
+  #   environment:
+  #     ASPNETCORE_URLS: "http://+:80"
+  #     <<: [*kestrel-cert, *router-microservice-defaults, *gc-light]
+  #     ConnectionStrings__Default: *postgres-connection
+  #     ConnectionStrings__Redis: "cache.stella-ops.local:6379"
+  #     Router__Enabled: "${DOCTOR_SCHEDULER_ROUTER_ENABLED:-true}"
+  #     Router__Messaging__ConsumerGroup: "doctor-scheduler"
+  #   volumes:
+  #     - *cert-volume
+  #   healthcheck:
+  #     test: ["CMD-SHELL", "bash -c 'echo > /dev/tcp/$(hostname)/80'"]
+  #     <<: *healthcheck-tcp
+  #   networks:
+  #     stellaops:
+  #       aliases:
+  #         - doctor-scheduler.stella-ops.local
+  #   labels: *release-labels
 
   # --- Slot 27: OpsMemory (src/AdvisoryAI/StellaOps.OpsMemory.WebService) ---
   opsmemory-web: