feat(apiproxy): settings/credentials/llm wire domains, frames, and write guard

Eight compiler-locked methods: settings.describe/update/replace serve
redacted layered namespace views (secrets structurally absent from every
layer, write-only in the update direction) and fold seam refusals into
settings-rejected; credentials.describe/set/unset expose value-free views
with credential-rejected on shadowed writes; llm.providers merges the
configurable directory with live routes and llm.models claims the
host-scoped catalog reservation through the buildModelCatalog extraction
session.models now shares. Three HostFrame invalidations bridge the seam
events (host/settings-changed, host/credentials-changed,
host/models-changed), and the connection route generalizes the native-
dialog check into a privileged-method set covering all four writes. The
fixture and both fake clients grow the same face.

packages/host/apiproxy/src/api/rpc.schema.ts

@@ -45,6 +45,8 @@ export const rpcErrorSchema: z.ZodType<RpcError> = z.discriminatedUnion('code',
   z.object({ code: z.literal('agent-busy'), message: z.string(), details: z.object({ reason: z.string() }) }),
   z.object({ code: z.literal('command-error'), message: z.string(), details: z.object({}) }),
   z.object({ code: z.literal('unknown-command'), message: z.string(), details: z.object({}) }),
+  z.object({ code: z.literal('settings-rejected'), message: z.string(), details: z.object({ ns: z.string() }) }),
+  z.object({ code: z.literal('credential-rejected'), message: z.string(), details: z.object({ ref: z.string() }) }),
   z.object({ code: z.literal('internal'), message: z.string(), details: z.object({}) }),
 ]) as unknown as z.ZodType<RpcError>