feat: plan agent v2, global chat mode, and user-mode v2 flag sync

- Add runtime/plan_agent_v2 package and shims; gateway/direct_loop/WS wiring

- Admin chat: interaction mode and specialist only in user menu; session API stores memory_mode and execution_mode only

- POST /admin/api/chat/user-mode mirrors plan_agent_version to AIA_EXPERT_PLAN_AGENT_V2_ENABLED (v2 to 1, v1 to 0)

- Composer cleanup (hidden mode select, no reasoning toggle in meta bar); tests and _local/system.env.example

Co-authored-by: Cursor <cursoragent@cursor.com>

tests/test_ws_gateway.py

@@ -295,7 +295,12 @@ class WsGatewayTests(unittest.TestCase):
                         "type": "req",
                         "id": "cs1",
                         "method": "chat.send",
-                        "params": {"sessionKey": "sess-chat", "message": "hi", "idempotencyKey": "idem-chat-1"},
+                        "params": {
+                            "sessionKey": "sess-chat",
+                            "message": "hi",
+                            "idempotencyKey": "idem-chat-1",
+                            "execution_mode": "plan",
+                        },
                     }
                 )
                 ack = None
@@ -309,6 +314,7 @@ class WsGatewayTests(unittest.TestCase):
                 assert ack.get("id") == "cs1"
                 assert ack.get("ok") is True
                 assert str((ack.get("payload") or {}).get("status") or "") == "started"
+                assert str((ack.get("payload") or {}).get("executionMode") or "") == "plan"
                 run_id = str((ack.get("payload") or {}).get("runId") or "")
                 assert run_id.strip() != ""