manaul correction attempts

src/conductor_tech_lead.py

@@ -5,7 +5,7 @@ This module implements the Tier 2 (Tech Lead) function for generating implementa
  It uses the LLM to analyze the track requirements and produce structured ticket definitions.
 
 Architecture:
- - Uses ai_client.send() for LLM communication
+ - Uses ai_client.send_result() for LLM communication
  - Uses mma_prompts.PROMPTS["tier2_sprint_planning"] for system prompt
  - Returns JSON array of ticket definitions
 

src/mcp_client.py

@@ -2370,7 +2370,7 @@ MCP_TOOL_SPECS: list[dict[str, Any]] = [
    "properties": {
     "target": {
      "type": "string",
-     "description": "Fully qualified name of the target (e.g., 'src.ai_client.send') or class.method.",
+     "description": "Fully qualified name of the target (e.g., 'src.ai_client.send_result') or class.method.",
     },
     "max_depth": {
      "type": "integer",

tests/test_ai_loop_regressions_20260614.py

@@ -46,7 +46,7 @@ def test_fr1_error_becomes_discussion_entry(mock_app: App, monkeypatch: pytest.M
  When send_result returns errors, _handle_request_event must enqueue a
  'response' event with status='error' and the error message in the text.
 
-  Currently broken: the code calls deprecated ai_client.send() which
+  Currently broken: the code calls deprecated ai_client.send_result() which
   silently returns '' on error. The empty string is then routed to the
   event_queue as a 'done' response and _on_comms_entry filters it out
   via `if text_content.strip():` (src/app_controller.py:3801).

tests/test_context_pruner.py

@@ -104,7 +104,7 @@ def test_token_reduction_logging(capsys):
  with pytest.MonkeyPatch().context() as m:
   m.setattr("builtins.open", lambda f, *args, **kwargs: type('obj', (object,), {'read': lambda s: code, '__enter__': lambda s: s, '__exit__': lambda s, *a: None})())
   m.setattr("pathlib.Path.exists", lambda s: True)
-  m.setattr("src.ai_client.send", lambda **kwargs: "DONE")
+  m.setattr("src.ai_client.send_result", lambda **kwargs: "DONE")
   
   run_worker_lifecycle(ticket, context, context_files=["test.py"])
  

tests/test_headless_verification.py

@@ -27,7 +27,7 @@ async def test_headless_verification_full_run(vlogger) -> None:
  vlogger.log_state("T2 Status Initial", "todo", t2.status)
 
  # We must patch where it is USED: multi_agent_conductor
- with patch("src.multi_agent_conductor.ai_client.send") as mock_send, \
+ with patch("src.multi_agent_conductor.ai_client.send_result") as mock_send, \
  patch("src.multi_agent_conductor.ai_client.reset_session") as mock_reset, \
  patch("src.multi_agent_conductor.confirm_spawn", return_value=(True, "mock_prompt", "mock_ctx")):
  # We need mock_send to return something that doesn't contain "BLOCKED"

tests/test_rag_integration.py

@@ -72,7 +72,7 @@ def test_rag_integration(mock_project):
  # message sent to the provider. We use 'wraps' to let the real logic run 
  # while still having a mock we can inspect. We also mock the internal 
  # _send_gemini which is what actually "sends to the provider".
- with patch('src.ai_client.send', wraps=ai_client.send) as mock_send:
+ with patch('src.ai_client.send_result', wraps=ai_client.send_result) as mock_send:
   with patch('src.ai_client._send_gemini') as mock_provider:
    mock_provider.return_value = "Mock AI Response"
    

tests/test_tiered_aggregation.py

@@ -40,7 +40,7 @@ def test_app_controller_do_generate_uses_persona_strategy(mock_build):
  assert call_kwargs.get("aggregation_strategy") == "full"
 
 @patch("src.summarize.summarise_file")
-@patch("src.multi_agent_conductor.ai_client.send")
+@patch("src.multi_agent_conductor.ai_client.send_result")
 def test_run_worker_lifecycle_uses_strategy(mock_send, mock_summarise, tmp_path):
  mock_send.return_value = "fake response"
  mock_summarise.return_value = "fake summary"