fix: add missing task_done() calls in agent runtime and other issues

python/packages/autogen-agentchat/src/autogen_agentchat/teams/_group_chat/_base_group_chat_manager.py

@@ -150,7 +150,8 @@ async def handle_agent_response(
             await self.update_message_thread(delta)
 
             # Remove the agent from the active speakers list.
-            self._active_speakers.remove(message.name)
+            if message.name in self._active_speakers:
+                self._active_speakers.remove(message.name)
             if len(self._active_speakers) > 0:
                 # If there are still active speakers, return without doing anything.
                 return

python/packages/autogen-core/src/autogen_core/_single_threaded_agent_runtime.py

@@ -707,6 +707,7 @@ async def _process_next(self) -> None:
                                 _warn_if_none(temp_message, "on_send")
                             except BaseException as e:
                                 future.set_exception(e)
+                                self._message_queue.task_done()
                                 return
                             if temp_message is DropMessage or isinstance(temp_message, DropMessage):
                                 event_logger.info(
@@ -718,6 +719,7 @@ async def _process_next(self) -> None:
                                     )
                                 )
                                 future.set_exception(MessageDroppedException())
+                                self._message_queue.task_done()
                                 return
 
                         message_envelope.message = temp_message
@@ -747,6 +749,7 @@ async def _process_next(self) -> None:
                             except BaseException as e:
                                 # TODO: we should raise the intervention exception to the publisher.
                                 logger.error(f"Exception raised in in intervention handler: {e}", exc_info=True)
+                                self._message_queue.task_done()
                                 return
                             if temp_message is DropMessage or isinstance(temp_message, DropMessage):
                                 event_logger.info(
@@ -757,6 +760,7 @@ async def _process_next(self) -> None:
                                         kind=MessageKind.PUBLISH,
                                     )
                                 )
+                                self._message_queue.task_done()
                                 return
 
                         message_envelope.message = temp_message
@@ -773,6 +777,7 @@ async def _process_next(self) -> None:
                         except BaseException as e:
                             # TODO: should we raise the exception to sender of the response instead?
                             future.set_exception(e)
+                            self._message_queue.task_done()
                             return
                         if temp_message is DropMessage or isinstance(temp_message, DropMessage):
                             event_logger.info(
@@ -784,6 +789,7 @@ async def _process_next(self) -> None:
                                 )
                             )
                             future.set_exception(MessageDroppedException())
+                            self._message_queue.task_done()
                             return
                         message_envelope.message = temp_message
                 task = asyncio.create_task(self._process_response(message_envelope))

python/packages/autogen-core/src/autogen_core/model_context/_token_limited_chat_completion_context.py

@@ -58,21 +58,25 @@ async def get_messages(self) -> List[LLMMessage]:
         """Get at most `token_limit` tokens in recent messages. If the token limit is not
         provided, then return as many messages as the remaining token allowed by the model client."""
         messages = list(self._messages)
+        trimmed = False
         if self._token_limit is None:
             remaining_tokens = self._model_client.remaining_tokens(messages, tools=self._tool_schema)
             while remaining_tokens < 0 and len(messages) > 0:
                 middle_index = len(messages) // 2
                 messages.pop(middle_index)
                 remaining_tokens = self._model_client.remaining_tokens(messages, tools=self._tool_schema)
+                trimmed = True
         else:
             token_count = self._model_client.count_tokens(messages, tools=self._tool_schema)
             while token_count > self._token_limit and len(messages) > 0:
                 middle_index = len(messages) // 2
                 messages.pop(middle_index)
                 token_count = self._model_client.count_tokens(messages, tools=self._tool_schema)
-        if messages and isinstance(messages[0], FunctionExecutionResultMessage):
-            # Handle the first message is a function call result message.
-            # Remove the first message from the list.
+                trimmed = True
+        if trimmed and messages and isinstance(messages[0], FunctionExecutionResultMessage):
+            # Only remove the function result if trimming actually occurred.
+            # Function call results are critical for tool-use flows and must be
+            # preserved when the context fits within the token limit.
             messages = messages[1:]
         return messages