feat: (maybe) stop response when new user message

If we get a new message before the LLM is done responding, interrupt it.

ref: N25B-452

src/control_backend/agents/bdi/bdi_core_agent.py

@@ -338,7 +338,7 @@ class BDICoreAgent(BaseAgent):
             yield
 
         @self.actions.add(".reply_with_goal", 3)
-        def _reply_with_goal(agent: "BDICoreAgent", term, intention):
+        def _reply_with_goal(agent, term, intention):
             """
             Let the LLM generate a response to a user's utterance with the current norms and a
             specific goal.

src/control_backend/agents/bdi/text_belief_extractor_agent.py

@@ -318,6 +318,9 @@ class TextBeliefExtractorAgent(BaseAgent):
                 async with httpx.AsyncClient() as client:
                     response = await client.post(
                         settings.llm_settings.local_llm_url,
+                        headers={"Authorization": f"Bearer {settings.llm_settings.api_key}"}
+                        if settings.llm_settings.api_key
+                        else {},
                         json={
                             "model": settings.llm_settings.local_llm_model,
                             "messages": [{"role": "user", "content": prompt}],