fix(agent): preserve Gemini thought_signature in MoA aggregator mode

When MoA mode is active with a Gemini model as the aggregator,
agent.model holds the virtual preset name (e.g. "closed"), not the
actual aggregator model name. The _sanitize_tool_calls_for_strict_api
call uses agent.model to decide whether to keep extra_content
(thought_signature) on tool_calls — since "closed" doesn't contain
"gemini", the thought_signature is stripped and the Gemini aggregator
rejects the next request with HTTP 400 (INVALID_ARGUMENT):
"Function call is missing a thought_signature in functionCall parts."

Fix: resolve the actual aggregator model name from moa_config
(conversation_loop) or last_aggregator_slot (chat_completion_helpers)
and pass it to _sanitize_tool_calls_for_strict_api so the
_model_consumes_thought_signature check sees the real Gemini model.

Closes #65092

agent/chat_completion_helpers.py

@@ -1946,7 +1946,17 @@ def handle_max_iterations(agent, messages: list, api_call_count: int) -> str:
             for internal_key in [k for k in api_msg if isinstance(k, str) and k.startswith("_")]:
                 api_msg.pop(internal_key, None)
             if _needs_sanitize:
-                agent._sanitize_tool_calls_for_strict_api(api_msg, model=agent.model)
+                # In MoA mode, agent.model is the virtual preset name,
+                # not the actual aggregator model.  Resolve the real
+                # aggregator model so Gemini preserves thought_signature.
+                _sanitize_model = agent.model
+                if agent.provider == "moa":
+                    _moa_client = getattr(agent, "client", None)
+                    if _moa_client is not None:
+                        _agg_slot = getattr(_moa_client, "last_aggregator_slot", None)
+                        if _agg_slot and _agg_slot.get("model"):
+                            _sanitize_model = _agg_slot["model"]
+                agent._sanitize_tool_calls_for_strict_api(api_msg, model=_sanitize_model)
             api_messages.append(api_msg)
 
         effective_system = agent._cached_system_prompt or ""

agent/conversation_loop.py

@@ -1103,7 +1103,16 @@ def run_conversation(
             # Uses new dicts so the internal messages list retains the fields
             # for Codex Responses compatibility.
             if agent._should_sanitize_tool_calls():
-                agent._sanitize_tool_calls_for_strict_api(api_msg, model=agent.model)
+                # In MoA mode, agent.model is the virtual preset name
+                # (e.g. "closed"), not the actual aggregator model.  Use
+                # the resolved aggregator model so Gemini aggregators
+                # correctly preserve thought_signature (extra_content).
+                _sanitize_model = agent.model
+                if agent.provider == "moa" and moa_config:
+                    _agg = moa_config.get("aggregator") or {}
+                    if _agg.get("model"):
+                        _sanitize_model = _agg["model"]
+                agent._sanitize_tool_calls_for_strict_api(api_msg, model=_sanitize_model)
             # Keep 'reasoning_details' - OpenRouter uses this for multi-turn reasoning context
             # The signature field helps maintain reasoning continuity
             api_messages.append(api_msg)