Add local search quality evaluation harness

Add an optional local evaluator that compares a baseline revision against a candidate checkout, computes deterministic stability metrics, and can call Gemini for judged ranking metrics when configured.

The harness isolates child runs with a temporary HOME and a node-free PATH so historical revisions cannot trigger Bird browser-cookie auth during evaluation.

Validation: uv run python -m unittest and local smoke/full deterministic eval runs.

tests/test_env_project.py

@@ -79,6 +79,23 @@ class TestConfigPrecedence(unittest.TestCase):
                 config = env.get_config()
                 self.assertEqual(config['BRAVE_API_KEY'], 'env-key')
 
+    def test_gemini_keys_load_from_project_env(self):
+        import tempfile
+        with tempfile.TemporaryDirectory() as tmpdir:
+            project_dir = Path(tmpdir) / ".claude"
+            project_dir.mkdir()
+            project_env = project_dir / "last30days.env"
+            project_env.write_text("GEMINI_API_KEY=gem-key\nGEMINI_MODEL=gemini-3-pro-preview\n")
+
+            with patch.object(Path, 'cwd', return_value=Path(tmpdir)), \
+                 patch.object(env, 'CONFIG_FILE', None), \
+                 patch.dict(os.environ, {}, clear=False):
+                os.environ.pop('GEMINI_API_KEY', None)
+                os.environ.pop('GEMINI_MODEL', None)
+                config = env.get_config()
+                self.assertEqual(config['GEMINI_API_KEY'], 'gem-key')
+                self.assertEqual(config['GEMINI_MODEL'], 'gemini-3-pro-preview')
+
 
 class TestConfigSource(unittest.TestCase):
     """Tests for _CONFIG_SOURCE tracking."""