chore: deprecate ToolResponseMessage in agent.resume API (#1566)

# Summary:
closes #1431 

# Test Plan:
LLAMA_STACK_CONFIG=fireworks pytest -s -v
tests/integration/agents/test_agents.py --safety-shield
meta-llama/Llama-Guard-3-8B --text-model
meta-llama/Llama-3.1-8B-Instruct

docs/_static/llama-stack-spec.html

@@ -9490,21 +9490,11 @@
                 "type": "object",
                 "properties": {
                     "tool_responses": {
-                        "oneOf": [
-                            {
-                                "type": "array",
-                                "items": {
-                                    "$ref": "#/components/schemas/ToolResponse"
-                                }
-                            },
-                            {
-                                "type": "array",
-                                "items": {
-                                    "$ref": "#/components/schemas/ToolResponseMessage"
-                                }
-                            }
-                        ],
-                        "description": "The tool call responses to resume the turn with. NOTE: ToolResponseMessage will be deprecated. Use ToolResponse."
+                        "type": "array",
+                        "items": {
+                            "$ref": "#/components/schemas/ToolResponse"
+                        },
+                        "description": "The tool call responses to resume the turn with."
                     },
                     "stream": {
                         "type": "boolean",

docs/_static/llama-stack-spec.yaml

@@ -6405,16 +6405,11 @@ components:
       type: object
       properties:
         tool_responses:
-          oneOf:
-            - type: array
-              items:
-                $ref: '#/components/schemas/ToolResponse'
-            - type: array
-              items:
-                $ref: '#/components/schemas/ToolResponseMessage'
+          type: array
+          items:
+            $ref: '#/components/schemas/ToolResponse'
           description: >-
-            The tool call responses to resume the turn with. NOTE: ToolResponseMessage
-            will be deprecated. Use ToolResponse.
+            The tool call responses to resume the turn with.
         stream:
           type: boolean
           description: Whether to stream the response.

llama_stack/apis/agents/agents.py

@@ -370,7 +370,7 @@ class AgentTurnResumeRequest(BaseModel):
     agent_id: str
     session_id: str
     turn_id: str
-    tool_responses: Union[List[ToolResponse], List[ToolResponseMessage]]
+    tool_responses: List[ToolResponse]
     stream: Optional[bool] = False
 
 
@@ -449,7 +449,7 @@ class Agents(Protocol):
         agent_id: str,
         session_id: str,
         turn_id: str,
-        tool_responses: Union[List[ToolResponse], List[ToolResponseMessage]],
+        tool_responses: List[ToolResponse],
         stream: Optional[bool] = False,
     ) -> Union[Turn, AsyncIterator[AgentTurnResponseStreamChunk]]:
         """Resume an agent turn with executed tool call responses.
@@ -460,7 +460,6 @@ class Agents(Protocol):
         :param session_id: The ID of the session to resume.
         :param turn_id: The ID of the turn to resume.
         :param tool_responses: The tool call responses to resume the turn with.
-            NOTE: ToolResponseMessage will be deprecated. Use ToolResponse.
         :param stream: Whether to stream the response.
         :returns: A Turn object if stream is False, otherwise an AsyncIterator of AgentTurnResponseStreamChunk objects.
         """

llama_stack/providers/inline/agents/meta_reference/agent_instance.py

@@ -218,18 +218,10 @@ class ChatAgent(ShieldRunnerMixin):
         steps = []
         messages = await self.get_messages_from_turns(turns)
         if is_resume:
-            if isinstance(request.tool_responses[0], ToolResponseMessage):
-                tool_response_messages = request.tool_responses
-                tool_responses = [
-                    ToolResponse(call_id=x.call_id, tool_name=x.tool_name, content=x.content)
-                    for x in request.tool_responses
-                ]
-            else:
-                tool_response_messages = [
-                    ToolResponseMessage(call_id=x.call_id, tool_name=x.tool_name, content=x.content)
-                    for x in request.tool_responses
-                ]
-                tool_responses = request.tool_responses
+            tool_response_messages = [
+                ToolResponseMessage(call_id=x.call_id, tool_name=x.tool_name, content=x.content)
+                for x in request.tool_responses
+            ]
             messages.extend(tool_response_messages)
             last_turn = turns[-1]
             last_turn_messages = self.turn_to_messages(last_turn)
@@ -252,7 +244,7 @@ class ChatAgent(ShieldRunnerMixin):
                 step_id=(in_progress_tool_call_step.step_id if in_progress_tool_call_step else str(uuid.uuid4())),
                 turn_id=request.turn_id,
                 tool_calls=(in_progress_tool_call_step.tool_calls if in_progress_tool_call_step else []),
-                tool_responses=tool_responses,
+                tool_responses=request.tool_responses,
                 completed_at=now,
                 started_at=(in_progress_tool_call_step.started_at if in_progress_tool_call_step else now),
             )

llama_stack/providers/inline/agents/meta_reference/agents.py

@@ -172,7 +172,7 @@ class MetaReferenceAgentsImpl(Agents):
         agent_id: str,
         session_id: str,
         turn_id: str,
-        tool_responses: Union[List[ToolResponse], List[ToolResponseMessage]],
+        tool_responses: List[ToolResponse],
         stream: Optional[bool] = False,
     ) -> AsyncGenerator:
         request = AgentTurnResumeRequest(