diff --git a/app/lingma_client.py b/app/lingma_client.py
index 93fa106..0e59011 100644
--- a/app/lingma_client.py
+++ b/app/lingma_client.py
@@ -896,15 +896,15 @@ class LingmaGatewayClient:
         is_reply: bool = False,
         tool_config: dict[str, Any] | None = None,
     ):
-        session_type = "developer" if ask_mode == "agent" else "chat"
+        session_type = "ask" if ask_mode == "agent" else "chat"
         payload = {
             "requestId": request_id,
             "sessionId": session_id,
             "sessionType": session_type,
-            "chatTask": "FREE_INPUT",
+            "chatTask": "chat" if ask_mode == "agent" else "FREE_INPUT",
             "mode": ask_mode,
             "stream": True,
-            "source": 1,
+            "source": "vscode" if ask_mode == "agent" else 1,
             "isReply": is_reply,
             "taskDefinitionType": "system",
             "content": prompt,
diff --git a/app/main.py b/app/main.py
index 4721929..6a429f2 100644
--- a/app/main.py
+++ b/app/main.py
@@ -504,13 +504,13 @@ async def v1_chat_completions(req: ChatCompletionsRequest, request: Request):
     #   1. Reuse the upstream sessionId so Lingma/Qwen hits its KV cache.
     #   2. Send only the new user message instead of the whole history.
     #   3. Stick the request to the pool instance that originally served it.
-    ask_mode = settings.default_ask_mode
-    if req.model.lower() in {"lingma-agent", "agent"}:
-        ask_mode = "agent"
-
     tool_config = _openai_tool_config(req)
     has_tooling_context = _openai_has_tooling_context(req, messages_dump)
 
+    ask_mode = settings.default_ask_mode
+    if req.model.lower() in {"lingma-agent", "agent"} or has_tooling_context:
+        ask_mode = "agent"
+
     reuse_eligible = (
         session_cache.enabled
         and ask_mode == "chat"
diff --git a/tests/test_tool_call_bridge.py b/tests/test_tool_call_bridge.py
index b10a5cb..bcf16dd 100644
--- a/tests/test_tool_call_bridge.py
+++ b/tests/test_tool_call_bridge.py
@@ -500,6 +500,7 @@ class ToolCallBridgeTests(unittest.IsolatedAsyncioTestCase):
         self.assertEqual(cfg["provider"], "openai")
         self.assertEqual(len(cfg["tools"]), 1)
         self.assertIsInstance(cfg["tool_choice"], dict)
+        self.assertEqual(spy_client.last_complete_args[2], "agent")
 
     async def test_openai_non_stream_does_not_forward_tool_config_when_disabled(self) -> None:
         spy_client = _SpyClient(stream_events=[], complete_result={"text": "ok", "toolEvents": []})
@@ -522,6 +523,7 @@ class ToolCallBridgeTests(unittest.IsolatedAsyncioTestCase):
 
         self.assertIn("tool_config", spy_client.last_complete_kwargs)
         self.assertIsNone(spy_client.last_complete_kwargs["tool_config"])
+        self.assertEqual(spy_client.last_complete_args[2], "agent")
 
 
     async def test_openai_tooling_context_disables_session_reuse_cache(self) -> None: