feat(secrets): wire the pull-only gRPC secrets hub — central hosts, sites sweep

scadaproj#3: the production secrets topology is central hosting a pull-only
gRPC hub with site nodes sweeping it. The SqlServer replicator stays in the
codebase and keeps working exactly as it did, but it is not the production
path — it needs every site node to hold a connection string to central's
database, which breaks ScadaBridge's standing rule that sites talk to central,
not to central's DB.

Selection is a new Secrets:Replication:Mode key alongside the existing
Secrets:Replication:Enabled flag. Absent or blank means SqlServer, so an
existing configuration that sets only Enabled behaves identically; an
unrecognised value is refused at startup naming the key, and refused whenever
it is present rather than only when replication is on — a typo should fail the
boot that introduced it, not some later boot that flips an unrelated flag.

Which HALF a node composes is a parameter, not a config key. Both composition
roots already know statically which they are (Program.cs is central,
SiteServiceRegistration is a site), and a role read from configuration is a
role that can be got wrong in the one direction that matters: a site hosting
the hub would serve central's whole secret inventory from inside the site
network to anything holding the shared token.

The hub is mapped onto the EXISTING central h2c control-plane listener
(ScadaBridge:Node:CentralGrpcPort, default 8083) beside CentralControlService
— the same listener and the same addressing convention sites already use, and
the same shape as the site's LocalDb sync endpoint sharing its gRPC port: two
disjoint service prefixes, two independent fail-closed gates. The
CentralControlAuthInterceptor on AddGrpc is prefix-scoped and passes hub calls
through; the hub's own SecretsHubAuthInterceptor, attached per-service by
AddZbSecretsGrpcHub, gates them on Secrets:GrpcHub:BearerToken.

Registration and mapping share one predicate (UsesGrpcHub) deliberately.
Mapping without registering would map a hub whose interceptor was never
attached — an anonymous endpoint serving every secret central holds, on a node
that looks completely healthy — so the two must not be able to drift.

gRPC mode fails CLOSED where SqlServer mode degrades: a missing endpoint or
bearer token is a startup failure, not a warning plus a local-only store. The
degraded outcome is precisely what the hub exists to prevent (a site quietly
serving secrets that never converge), and the package's own errors name the
exact key, so they are left unwrapped.

Templates are default-OFF: Enabled stays false, Mode stays SqlServer, and
Secrets:GrpcHub ships with an empty BearerToken and Endpoint. Empty is
fail-closed, not open. The token is documented as appsettings/env only — never
a ${secret:} reference, since resolving one is what the hub exists to make
possible — the same bootstrap rule the mesh pre-shared keys follow.

Known asymmetry, recorded in the template: CentralGrpcEndpoints is a LIST that
fails over across the central pair, but the hub client dials a SINGLE endpoint,
so a sweep against a downed central node stalls instead of failing over. That
is survivable — the sweep is best-effort and the site keeps serving its full
local last-known-good store — but secrets stop converging until that node is
back.

Claude-Session: https://claude.ai/code/session_014WNM4vjoVksyyBraTXSZE1

src/ZB.MOM.WW.ScadaBridge.Host/appsettings.json

@@ -14,13 +14,28 @@
     "MasterKey": { "Source": "Environment", "EnvVarName": "ZB_SECRETS_MASTER_KEY" },
     "RunMigrationsOnStartup": true,
     "ResolveCacheTtl": "00:00:30",
-    "Replication": { "Enabled": false },
+    "Replication": {
+      "_comment": "Default-OFF pin: no deployment enables clustered secret replication by shipping this file. Mode selects the transport when Enabled is true - 'SqlServer' (the default, kept so an existing config that sets only Enabled behaves exactly as before) or 'Grpc' (the PRODUCTION topology, scadaproj#3: central hosts a pull-only hub under Secrets:GrpcHub and sites sweep it, so a site never needs a connection string to central's database). Every node in either topology must carry the SAME ZB_SECRETS_MASTER_KEY - only ciphertext crosses the wire, so a node with a different KEK fails closed on resolve with a kek_id mismatch that reads like corruption but is a deployment error.",
+      "Enabled": false,
+      "Mode": "SqlServer"
+    },
     "SqlServer": {
       "ConnectionString": "",
       "SchemaName": "zbsecrets",
       "SyncInterval": "00:00:30",
       "SyncOnStartup": true,
-      "_comment": "Hub connstr must be seeded in the LOCAL store (or supplied via env) - it cannot come from the hub itself. Enable via Secrets:Replication:Enabled."
+      "_comment": "Hub connstr must be seeded in the LOCAL store (or supplied via env) - it cannot come from the hub itself. Enable via Secrets:Replication:Enabled + Mode=SqlServer."
+    },
+    "GrpcHub": {
+      "_comment": "Read ONLY when Secrets:Replication:Enabled is true AND Mode=Grpc; inert otherwise. ONE section, both halves: CENTRAL reads BearerToken + MaxNamesPerRequest and hosts the hub on its CentralGrpcPort h2c listener (default 8083, alongside CentralControlService); a SITE reads Endpoint + BearerToken + the sweep timings and pulls. Replication is pull-only by wire contract - the proto has no write RPC - so secrets originate at central and a site cannot push. FAIL-CLOSED: unlike SqlServer mode there is no local-only fallback; a missing BearerToken (either role) or Endpoint (site) is a startup failure naming the key.",
+      "_bearerToken": "Shared credential every follower presents. Supply it from appsettings or the environment (Secrets__GrpcHub__BearerToken), NOT as a ${secret:...} reference - resolving that reference is what the hub exists to make possible, so it cannot come from the hub. Same rationale and same handling as the mesh pre-shared keys. Never commit a real value here; the empty default below is fail-closed, not open.",
+      "BearerToken": "",
+      "_endpoint": "SITE ONLY. Absolute http/https URI of a CENTRAL node's gRPC (h2c) port - the same address family as ScadaBridge:Communication:CentralGrpcEndpoints, e.g. 'http://central-a-host:8083'. NOT via Traefik (HTTP/1 only). NOTE the asymmetry with CentralGrpcEndpoints: that is a LIST and fails over across the central pair, whereas the hub client dials a SINGLE endpoint. A sweep against a downed central-a therefore stalls rather than failing over - which is survivable because the sweep is best-effort (one warning per interval, then retry) and the site keeps serving its full local last-known-good store, but it does mean secrets stop converging until that central node returns.",
+      "Endpoint": "",
+      "SyncInterval": "00:00:30",
+      "SyncOnStartup": true,
+      "CallDeadline": "00:00:30",
+      "MaxNamesPerRequest": 1000
     }
   },
   "Serilog": {