fix(sandbox): improve sandbox security and preserve multimodal content (#2114)

* fix: improve sandbox security and preserve multimodal content

* Add unit test modifications for test_injects_uploaded_files_tag_into_list_content

* format updated_content

* Add regression tests for multimodal upload content and host bash default safety

backend/packages/harness/deerflow/agents/middlewares/uploads_middleware.py

@@ -262,21 +262,25 @@ class UploadsMiddleware(AgentMiddleware[UploadsMiddlewareState]):
         files_message = self._create_files_message(new_files, historical_files)
 
         # Extract original content - handle both string and list formats
-        original_content = ""
-        if isinstance(last_message.content, str):
-            original_content = last_message.content
-        elif isinstance(last_message.content, list):
-            text_parts = []
-            for block in last_message.content:
-                if isinstance(block, dict) and block.get("type") == "text":
-                    text_parts.append(block.get("text", ""))
-            original_content = "\n".join(text_parts)
+        original_content = last_message.content
+        if isinstance(original_content, str):
+            # Simple case: string content, just prepend files message
+            updated_content = f"{files_message}\n\n{original_content}"
+        elif isinstance(original_content, list):
+            # Complex case: list content (multimodal), preserve all blocks
+            # Prepend files message as the first text block
+            files_block = {"type": "text", "text": f"{files_message}\n\n"}
+            # Keep all original blocks (including images)
+            updated_content = [files_block, *original_content]
+        else:
+            # Other types, preserve as-is
+            updated_content = original_content
 
         # Create new message with combined content.
         # Preserve additional_kwargs (including files metadata) so the frontend
         # can read structured file info from the streamed message.
         updated_message = HumanMessage(
-            content=f"{files_message}\n\n{original_content}",
+            content=updated_content,
             id=last_message.id,
             additional_kwargs=last_message.additional_kwargs,
         )

backend/packages/harness/deerflow/sandbox/security.py

@@ -39,7 +39,7 @@ def is_host_bash_allowed(config=None) -> bool:
 
     sandbox_cfg = getattr(config, "sandbox", None)
     if sandbox_cfg is None:
-        return True
+        return False
     if not uses_local_sandbox_provider(config):
         return True
     return bool(getattr(sandbox_cfg, "allow_host_bash", False))

backend/tests/test_local_bash_tool_loading.py

@@ -1,5 +1,6 @@
 from types import SimpleNamespace
 
+from deerflow.sandbox.security import is_host_bash_allowed
 from deerflow.tools.tools import get_available_tools
 
 
@@ -79,3 +80,8 @@ def test_get_available_tools_keeps_bash_for_aio_sandbox(monkeypatch):
 
     assert "bash" in names
     assert "ls" in names
+
+
+def test_is_host_bash_allowed_defaults_false_when_sandbox_missing():
+    assert is_host_bash_allowed(SimpleNamespace()) is False
+    assert is_host_bash_allowed(SimpleNamespace(sandbox=None)) is False

backend/tests/test_uploads_middleware_core_logic.py

@@ -256,8 +256,10 @@ class TestBeforeAgent:
 
         assert result is not None
         updated_msg = result["messages"][-1]
-        assert "<uploaded_files>" in updated_msg.content
-        assert "analyse this" in updated_msg.content
+        assert isinstance(updated_msg.content, list)
+        combined_text = "\n".join(block.get("text", "") for block in updated_msg.content if isinstance(block, dict))
+        assert "<uploaded_files>" in combined_text
+        assert "analyse this" in combined_text
 
     def test_preserves_additional_kwargs_on_updated_message(self, tmp_path):
         mw = _middleware(tmp_path)