refactor(shared): split model-availability.ts into model resolution modules

Extract model availability checking pipeline:
- available-models-fetcher.ts: top-level model fetching orchestration
- model-cache-availability.ts, models-json-cache-reader.ts
- provider-models-cache-model-reader.ts: provider cache reading with null guard
- fallback-model-availability.ts, model-name-matcher.ts
- open-code-client-accessors.ts, open-code-client-shapes.ts
- record-type-guard.ts

src/shared/model-cache-availability.ts

@@ -0,0 +1,14 @@
+import { existsSync } from "fs"
+import { join } from "path"
+import { getOpenCodeCacheDir } from "./data-path"
+import { hasProviderModelsCache } from "./connected-providers-cache"
+
+export function __resetModelCache(): void {}
+
+export function isModelCacheAvailable(): boolean {
+	if (hasProviderModelsCache()) {
+		return true
+	}
+	const cacheFile = join(getOpenCodeCacheDir(), "models.json")
+	return existsSync(cacheFile)
+}