fix(guardrails): scope the logging_only response scan once

Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>

litellm/integrations/custom_guardrail.py

@@ -960,12 +960,14 @@ class CustomGuardrail(CustomLogger):
 
     def _chat_shaped_request(
         self,
-        scratch_request: dict,  # mutable-ok: CustomLogger.async_logging_hook contract
+        scratch_request: Mapping[str, object],
         translation: "BaseTranslation",
-    ) -> dict:  # mutable-ok: BaseTranslation.process_output_response contract
+    ) -> dict[str, object]:  # mutable-ok: BaseTranslation.process_output_response contract
         """The logged request in OpenAI chat shape, for an output scan whose translation differs from the input's."""
-        context: Final = translation.request_scan_context(scratch_request, self)
+        messages, tools = translation.chat_shaped_request_conversation(
-        return {**scratch_request, "messages": list(context.structured_messages), "tools": list(context.tools)}
+            dict(scratch_request)  # mutable-ok: BaseTranslation.chat_shaped_request_conversation requires a dict
+        )
+        return {**scratch_request, "messages": list(messages), "tools": list(tools)}
 
     def supports_scan_only_tool_results(self) -> bool:
         """Whether this guardrail can scan tool-result content.

litellm/llms/anthropic/chat/guardrail_translation/handler.py

@@ -528,23 +528,26 @@ class AnthropicMessagesHandler(BaseTranslation):
         )
         return result if result else None
 
-    def request_scan_context(self, data: dict, guardrail_to_apply: "CustomGuardrail") -> RequestScanContext:
+    def chat_shaped_request_conversation(
+        self, data: dict
+    ) -> tuple[tuple[AllMessageValues, ...], tuple[ChatCompletionToolParam, ...]]:
         if data.get("messages") is None:
-            return RequestScanContext()
+            return (), ()
         translated: Final = self._translate_to_openai(
             {key: value for key, value in data.items() if key != "system"}  # mutable-ok: API message payload
         )
-        hoisted_system_message: Final = (
+        hoisted_system_message: Final = self._hoisted_top_level_system_message(data)
-            None
+        messages: Final = (
-            if effective_skip_system_message_for_guardrail(guardrail_to_apply)
+            *(() if hoisted_system_message is None else (hoisted_system_message,)),
-            else self._hoisted_top_level_system_message(data)
+            *translated["messages"],
-        )
-        return RequestScanContext.scoped(
-            (*(() if hoisted_system_message is None else (hoisted_system_message,)), *translated["messages"]),
-            tuple(tool for tool in translated.get("tools") or () if not is_provider_native_tool_dict(tool)),
-            guardrail_to_apply,
-            skip_system=False,
         )
+        tools: Final = tuple(tool for tool in translated.get("tools") or () if not is_provider_native_tool_dict(tool))
+        return messages, tools
+
+    def request_scan_context(self, data: dict, guardrail_to_apply: "CustomGuardrail") -> RequestScanContext:
+        if data.get("messages") is None:
+            return RequestScanContext()
+        return RequestScanContext.scoped(*self.chat_shaped_request_conversation(data), guardrail_to_apply)
 
     async def process_input_messages(
         self,

litellm/llms/base_llm/guardrail_translation/base_translation.py

@@ -298,11 +298,16 @@ class BaseTranslation(ABC):
         """
         return None
 
+    def chat_shaped_request_conversation(
+        self, data: dict
+    ) -> tuple[tuple["AllMessageValues", ...], tuple["ChatCompletionToolParam", ...]]:
+        """The full, unscoped request turns and tool definitions in OpenAI chat shape."""
+        return tuple(self.get_structured_messages(data) or ()), tuple(data.get("tools") or ())
+
     def request_scan_context(self, data: dict, guardrail_to_apply: "CustomGuardrail") -> RequestScanContext:
         """Override wherever ``process_input_messages`` scopes or translates the request differently."""
-        return RequestScanContext.scoped(
+        messages, tools = self.chat_shaped_request_conversation(data)
-            self.get_structured_messages(data) or (), data.get("tools") or (), guardrail_to_apply
+        return RequestScanContext.scoped(messages, tools, guardrail_to_apply)
-        )
 
     def with_response_context(
         self,

tests/test_litellm/integrations/test_custom_guardrail.py

@@ -2699,6 +2699,22 @@ class TestLoggingOnlyApplyGuardrail:
             ("response", [*expected_request, {"role": "assistant", "content": "general kenobi"}], expected_tools),
         ]
 
+    @pytest.mark.asyncio
+    async def test_anthropic_messages_response_scan_keeps_reply_when_scoping_empties_request(self):
+        class _ContextObserver(_ApplyOnlyObserver):
+            @log_guardrail_information
+            async def apply_guardrail(self, inputs, request_data, input_type, logging_obj=None):
+                self.calls.append((input_type, inputs.get("structured_messages"), inputs.get("tools")))
+                return inputs
+
+        guardrail = _ContextObserver()
+        guardrail.scan_only_tool_results = True
+        kwargs, response = _logged_call([{"role": "user", "content": "What is the capital of France?"}])
+
+        await guardrail.async_logging_hook(kwargs, response, CallTypes.anthropic_messages.value)
+
+        assert guardrail.calls == [("response", [{"role": "assistant", "content": "general kenobi"}], None)]
+
     @pytest.mark.asyncio
     async def test_async_success_handler_records_verdict_in_standard_logging_object(self):
         import datetime as dt