feat: add dynamic variables and startup result handling in batch run schemas

- Introduced dynamic_variables field in BatchRunCreate schema to support flexible input types.
- Added BatchStartupResult schema to encapsulate startup state, including greeting and auto-confirmed messages.
- Updated BatchRunCaseOut schema to include startup results, enhancing the structure for test run outputs.
- Enhanced test run orchestration to prepare and validate dynamic configurations, ensuring isolated variable management across test cases.
- Implemented changes in the frontend to accommodate new data structures for improved test case execution and result tracking.

backend/services/brains/workflow_brain.py

@@ -951,14 +951,9 @@ class WorkflowBrain(BaseBrain):
         else:
             raise ValueError(f"系统工具 {tool.id} 缺少有效 kind")
 
-        runtime = self._require_runtime()
-        if runtime.tool_executor_factory is None:
-            return schema
-
-        async def mock_system_tool(args, _flow_manager):
-            return await self._tools.execute(tool, dict(args or {}))
-
-        return replace(schema, handler=mock_system_tool)
+        # Platform-owned system behavior is deterministic. Keep it real in all
+        # runtimes while an injected executor isolates only external tools.
+        return schema
 
     def _workflow_update_state_tool(
         self,