feat: skip model resolution for delegated tasks when provider cache not yet created

Before provider cache exists (first run), resolveModelForDelegateTask now
returns undefined instead of guessing a model. This lets OpenCode use its
system default model when no model is specified in the prompt body.

User-specified model overrides still take priority regardless of cache state.

src/tools/delegate-task/model-selection.ts

@@ -2,6 +2,7 @@ import type { FallbackEntry } from "../../shared/model-requirements"
 import { normalizeModel } from "../../shared/model-normalization"
 import { fuzzyMatchModel } from "../../shared/model-availability"
 import { transformModelForProvider } from "../../shared/provider-model-id-transform"
+import { hasConnectedProvidersCache, hasProviderModelsCache } from "../../shared/connected-providers-cache"
 
 function isExplicitHighModel(model: string): boolean {
   return /(?:^|\/)[^/]+-high$/.test(model)
@@ -25,6 +26,12 @@ export function resolveModelForDelegateTask(input: {
     return { model: userModel }
   }
 
+  // Before provider cache is created (first run), skip model resolution entirely.
+  // OpenCode will use its system default model when no model is specified in the prompt.
+  if (input.availableModels.size === 0 && !hasProviderModelsCache() && !hasConnectedProvidersCache()) {
+    return undefined
+  }
+
   const categoryDefault = normalizeModel(input.categoryDefaultModel)
   const explicitHighBaseModel = categoryDefault ? getExplicitHighBaseModel(categoryDefault) : null
   const explicitHighModel = explicitHighBaseModel ? categoryDefault : undefined