step-3: add router refactor with streaming support (orchestrator.py)

- orchestrate_v3(user_id, message, context): classifies intent, returns
  (agent_name, agent_instance) — caller drives execution
- orchestrate_v3_stream(user_id, message, context): yields (agent_name, token)
  pairs; first yield is always (agent_name, "") as a domain-detection signal
- ChatAgent.handle_stream(): default implementation yields handle() result as
  one chunk; subclasses override for true token-level streaming
- Fix stale test_orchestrator.py assertions that expected a JSON final frame
  that orchestrate_stream never emitted

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

tests/test_orchestrator.py

@@ -302,7 +302,7 @@ class TestOrchestrateStream:
         assert len(chunks) >= 1
 
     @pytest.mark.asyncio
-    async def test_last_chunk_is_final_json_frame(
+    async def test_all_chunks_are_plain_text(
         self, reg: AgentRegistry
     ) -> None:
         with patch("app.core.orchestrator._make_llm") as mock_cls:
@@ -310,13 +310,12 @@ class TestOrchestrateStream:
             request = ChatRequest(message="add a task", execution_mode="direct")
             chunks = [chunk async for chunk in orchestrate_stream(request, reg)]
 
-        last = json.loads(chunks[-1])
-        assert last["done"] is True
-        assert "response" in last
-        assert "actions" in last
+        # orchestrate_stream yields plain text chunks only — no JSON final frame
+        for chunk in chunks:
+            assert isinstance(chunk, str)
 
     @pytest.mark.asyncio
-    async def test_final_frame_response_matches_agent_output(
+    async def test_concatenated_chunks_equal_full_response(
         self, reg: AgentRegistry
     ) -> None:
         with patch("app.core.orchestrator._make_llm") as mock_cls:
@@ -324,8 +323,8 @@ class TestOrchestrateStream:
             request = ChatRequest(message="create a task", execution_mode="direct")
             chunks = [chunk async for chunk in orchestrate_stream(request, reg)]
 
-        final = json.loads(chunks[-1])
-        assert final["response"] == "task: create a task"
+        full_text = "".join(chunks)
+        assert full_text == "task: create a task"
 
     @pytest.mark.asyncio
     async def test_text_chunks_before_final_frame(