fix(conductor): Apply review suggestions for track 'live_gui_testing_20260223'

tests/test_agent_tools_wiring.py

@@ -8,16 +8,15 @@ sys.path.append(os.path.abspath(os.path.join(os.path.dirname(__file__), "..")))
 
 from ai_client import set_agent_tools, _build_anthropic_tools
 
-def test_set_agent_tools_gemini():
+def test_set_agent_tools():
-    with patch('ai_client._ensure_gemini_client'):
+    # Correct usage: pass a dict
-        set_agent_tools('gemini', ['read_file', 'list_directory'])
+    agent_tools = {"read_file": True, "list_directory": False}
-        # Implementation details check would go here
+    set_agent_tools(agent_tools)
 
 def test_build_anthropic_tools_conversion():
-    # Test that MCP tools are correctly formatted for Anthropic
+    # _build_anthropic_tools takes no arguments and uses the global _agent_tools
-    mcp_tools = [
+    # We set a tool to True and check if it appears in the output
-        {"name": "test_tool", "description": "desc", "input_schema": {"type": "object", "properties": {}}}
+    set_agent_tools({"read_file": True})
-    ]
+    anthropic_tools = _build_anthropic_tools()
-    anthropic_tools = _build_anthropic_tools(mcp_tools)
+    tool_names = [t["name"] for t in anthropic_tools]
-    assert len(anthropic_tools) == 1
+    assert "read_file" in tool_names
-    assert anthropic_tools[0]['name'] == 'test_tool'

tests/test_history_bleed.py

@@ -22,4 +22,5 @@ def test_get_history_bleed_stats_basic():
     stats = ai_client.get_history_bleed_stats()
     assert 'current' in stats
     assert 'limit' in stats
-    assert stats['limit'] == 1000
+    # ai_client.py hardcodes Gemini limit to 900_000
+    assert stats['limit'] == 900000

tests/test_mcp_perf_tool.py

@@ -10,10 +10,10 @@ import mcp_client
 
 def test_mcp_perf_tool_retrieval():
     # Test that the MCP tool can call performance_monitor metrics
-    mock_app = MagicMock()
+    mock_metrics = {"fps": 60, "last_frame_time_ms": 16.6}
-    mock_app.perf_monitor.get_metrics.return_value = {"fps": 60}
     
-    # Simulate tool call
+    # Simulate tool call by patching the callback
-    with patch('mcp_client.get_app_instance', return_value=mock_app):
+    with patch('mcp_client.perf_monitor_callback', return_value=mock_metrics):
-        # We assume there's a tool named 'get_performance_metrics' in the MCP client
+        result = mcp_client.get_ui_performance()
-        pass
+        assert "60" in result
+        assert "16.6" in result