Build and run agents you can see, understand and trust.
Python agentchatbotlarge-language-modelsllmllm-agentmcpmulti-agentmulti-modalreact-agentrealtime-agentvoice-agentvoice-assistant
feat(agent): filter thinking content before broadcast (#1332)
qbc346dce99a514
diff --git a/src/agentscope/agent/_agent_base.py b/src/agentscope/agent/_agent_base.pyindex 7a49b2702..9184dda45 100644--- a/src/agentscope/agent/_agent_base.py+++ b/src/agentscope/agent/_agent_base.py@@ -470,10 +470,48 @@ class AgentBase(StateModule, metaclass=_AgentMeta):self,msg: Msg | list[Msg] | None,) -> None:- """Broadcast the message to all subscribers."""+ """Broadcast the message to all subscribers.++ Thinking blocks are stripped before broadcasting, since they represent+ the agent's internal reasoning and should not be visible to others.+ """+ if msg is None:+ return++ broadcast_msg = self._strip_thinking_blocks(msg)+for subscribers in self._subscribers.values():for subscriber in subscribers:- await subscriber.observe(msg)+ await subscriber.observe(broadcast_msg)++ @staticmethod+ def _strip_thinking_blocks(msg: Msg | list[Msg]) -> Msg | list[Msg]:+ """Remove thinking blocks from message(s) before sharing with other+ agents."""+ if isinstance(msg, list):+ return [AgentBase._strip_thinking_blocks_single(m) for m in msg]+ return AgentBase._strip_thinking_blocks_single(msg)++ @staticmethod+ def _strip_thinking_blocks_single(msg: Msg) -> Msg:+ """Remove thinking blocks from a single message."""+ if not isinstance(msg.content, list):+ return msg++ filtered = [b for b in msg.content if b.get("type") != "thinking"]+ if len(filtered) == len(msg.content):+ return msg++ new_msg = Msg(+ name=msg.name,+ content=filtered,+ role=msg.role,+ metadata=msg.metadata,+ timestamp=msg.timestamp,+ invocation_id=msg.invocation_id,+ )+ new_msg.id = msg.id+ return new_msgasync def handle_interrupt(self,