fix: use correct google-genai async API and remove debug endpoint

The google-genai SDK uses `client.aio.models.generate_content()` for
async calls, not `client.models.generate_content_async()` which doesn't
exist. Also removes the temporary /ai/provider-debug endpoint.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

backend/app/api/endpoints/ai_builder.py

@@ -54,34 +54,6 @@ def _require_ai_enabled() -> None:
             detail="AI flow builder is not configured. Set GOOGLE_AI_API_KEY or ANTHROPIC_API_KEY.",
         )
 
-
-@router.get("/provider-debug")
-async def provider_debug(
-    current_user: Annotated[User, Depends(get_current_active_user)],
-    _: None = Depends(require_engineer_or_admin),
-):
-    """Temporary debug endpoint — shows which AI provider would be selected."""
-    from app.core.ai_provider import get_ai_provider
-    has_gemini_key = bool(settings.GOOGLE_AI_API_KEY)
-    has_anthropic_key = bool(settings.ANTHROPIC_API_KEY)
-    provider_setting = settings.AI_PROVIDER
-    try:
-        provider = get_ai_provider()
-        provider_type = type(provider).__name__
-    except RuntimeError as e:
-        provider_type = f"ERROR: {e}"
-    return {
-        "ai_provider_setting": provider_setting,
-        "has_gemini_key": has_gemini_key,
-        "gemini_key_prefix": settings.GOOGLE_AI_API_KEY[:8] + "..." if settings.GOOGLE_AI_API_KEY else None,
-        "has_anthropic_key": has_anthropic_key,
-        "anthropic_key_prefix": settings.ANTHROPIC_API_KEY[:8] + "..." if settings.ANTHROPIC_API_KEY else None,
-        "selected_provider": provider_type,
-        "gemini_model": settings.AI_MODEL_GEMINI,
-        "anthropic_model": settings.AI_MODEL_ANTHROPIC,
-    }
-
-
 @router.get("/quota", response_model=AIQuotaStatusResponse)
 async def get_quota(
     current_user: Annotated[User, Depends(get_current_active_user)],

backend/app/core/ai_provider.py

@@ -68,7 +68,7 @@ class GeminiProvider(AIProvider):
             response_mime_type="application/json",
         )
 
-        response = await client.models.generate_content_async(
+        response = await client.aio.models.generate_content(
             model=self._model,
             contents=contents,
             config=config,

backend/tests/test_ai_provider.py

@@ -137,7 +137,7 @@ class TestGeminiProvider:
         mock_response.usage_metadata = mock_usage
 
         mock_client = MagicMock()
-        mock_client.models.generate_content_async = AsyncMock(
+        mock_client.aio.models.generate_content = AsyncMock(
             return_value=mock_response
         )
 
@@ -172,7 +172,7 @@ class TestGeminiProvider:
         assert input_tokens == 80
         assert output_tokens == 40
 
-        mock_client.models.generate_content_async.assert_called_once()
+        mock_client.aio.models.generate_content.assert_called_once()
 
     @pytest.mark.asyncio
     async def test_generate_json_handles_none_usage(self):
@@ -185,7 +185,7 @@ class TestGeminiProvider:
         mock_response.usage_metadata = mock_usage
 
         mock_client = MagicMock()
-        mock_client.models.generate_content_async = AsyncMock(
+        mock_client.aio.models.generate_content = AsyncMock(
             return_value=mock_response
         )