How to avoid 400 internal error while developing multi-agents application

[ziyilin]

I'm developing a multi-agents application. There is a master agent that runs the workflow according to the predefined system message, and calls other agent to do certain sub-task. The code would be like:

public class MasterAgent {
@SessionScoped
@RegisterAiService
public interface MasterAgent {
    @SystemMessage("""
            You're an expert of .... 
            You can call subtaskagent to finish the subtask with the collected information.
            ...
            """)
    @ToolBox(SubTaskAgent.class)
    Multi<String> chat(String userMessage);
}

public class SubTaskAgent{
@SessionScoped
@RegisterAiService
public interface ConflictSolverAgent {

    @SystemMessage("""
     You're an expert of ... . Please summarize the {{information}}.
     ...
""")
    @Tool(name="subtaskagent", value = "...")
    String chat(@V("information") String information);
}

The application could work well sometimes, but it often failed with 400 error such as:

Response:
- status code: 400
- headers: [vary: Origin,Access-Control-Request-Method,Access-Control-Request-Headers, Accept-Encoding], [x-request-id: 25440da5-9cb7-4bcb-bfaf-f4d393e9b01d], [x-dashscope-call-gateway: true], [content-type: application/json], [content-length: 436], [req-cost-time: 526], [req-arrive-time: 1761564475982], [resp-start-time: 1761564476509], [x-envoy-upstream-service-time: 522], [set-cookie: acw_tc=25440da5-9cb7-4bcb-bfaf-f4d393e9b01dc59610161239f7dde1add1d1207eb086;path=/;HttpOnly;Max-Age=1800], [date: Mon, 27 Oct 2025 11:27:56 GMT], [server: istio-envoy]
- body: {"error":{"code":"invalid_parameter_error","param":null,"message":"<400> InternalError.Algo.InvalidParameter: An assistant message with \"tool_calls\" must be followed by tool messages responding to each \"tool_call_id\". The following tool_call_ids did not have response messages: message[3].role","type":"invalid_request_error"},"id":"chatcmpl-25440da5-9cb7-4bcb-bfaf-f4d393e9b01d","request_id":"25440da5-9cb7-4bcb-bfaf-f4d393e9b01d"}

How can I prevent such failure?

[geoand]

Thanks for reporting!

This could be an issue with memory, but without seeing it in action, its hard to tell. Would you be able to upload a sample application that exhibits the behavior you describe?

Thanks

[ziyilin]

Thanks for replying. I found this issue not only exists when AI to AI, but happens in any tool calls. I can't provide the sample application now. But here is the execution log:

Request:
- method: POST
- url: https://dashscope.aliyuncs.com/compatible-mode/v1/chat/completions
- headers: [Accept: text/event-stream], [Authorization: Be...7f], [Content-Type: application/json], [User-Agent: langchain4j-openai], [content-length: 58213]
- body: {
  "model" : "qwen3-max",
  "messages" : [ {
    "role" : "system",
    "content" : "..."
  }, {
    "role" : "assistant",
    "content" : "...\n\n",
    "tool_calls" : [ {
      "id" : "call_c252712692be482389df7921",
      "type" : "function",
      "function" : {
        "name" : "executeCommandLine",
        "arguments" : "{\"command\": \"git diff substratevm/src/com.oracle.svm.hosted/src/com/oracle/svm/hosted/NativeImageClassLoaderSupport.java\", \"workingDirectory\": \"..."}"
      }
    } ]
  }, {
    "role" : "tool",
    "tool_call_id" : "call_c252712692be482389df7921",
    "content" : "\"Exit code:0\\ndiff ..."
  }, {
    "role" : "assistant",
    "content" : "...：\n\n",
    "tool_calls" : [ {
      "id" : "call_0877b1fbc7d940e29dde044d",
      "type" : "function",
      "function" : {
        "name" : "executeCommandLine",
        "arguments" : "{\"command\": \"git show a760a3d5171 -- substratevm/src/com.oracle.svm.hosted/src/com/oracle/svm/hosted/NativeImageClassLoaderSupport.java\", \"workingDirectory\": \"..."}"
      }
    } ]
  }, {
    "role" : "tool",
    "tool_call_id" : "call_0877b1fbc7d940e29dde044d",
    "content" : "\"Exit code:0 ..."
  }, {
    "role" : "assistant",
    "content" : "...\n\n",
    "tool_calls" : [ {
      "id" : "call_2df3eb6ba05644cfbaeadb99",
      "type" : "function",
      "function" : {
        "name" : "executeCommandLine",
        "arguments" : "{\"command\": \"git log --oneline -L 74,+13:substratevm/src/com.oracle.svm.hosted/src/com/oracle/svm/hosted/NativeImageClassLoaderSupport.java\", \"workingDirectory\": \"..."}"
      }
    } ]
  }, {
    "role" : "assistant",
    "content" : "...\n\n",
    "tool_calls" : [ {
      "id" : "call_a98e40bd31184d08b723b2c2",
      "type" : "function",
      "function" : {
        "name" : "executeCommandLine",
        "arguments" : "{\"command\": \"git show a760a3d5171 -- substratevm/src/com.oracle.svm.hosted/src/com/oracle/svm/hosted/NativeImageClassLoaderSupport.java\", \"workingDirectory\": \"...\"}"
      }
    } ]
  }, {
    "role" : "tool",
    "tool_call_id" : "call_a98e40bd31184d08b723b2c2",
    "content" : "\"Exit code:0 ..."
  }, {
    "role" : "tool",
    "tool_call_id" : "call_2df3eb6ba05644cfbaeadb99",
    "content" : "\"Exit code:0 ..."
  } ],
  "temperature" : 0.3,
  "top_p" : 1.0,
  "stream" : true,
  "stream_options" : {
    "include_usage" : true
  },
  "max_tokens" : 4000,
  "presence_penalty" : 0.0,
  "frequency_penalty" : 2.0,
  "tools" : [ {
    "type" : "function",
    "function" : {
      "name" : "executeCommandLine",
      "description" : "Execute command in another process",
      "parameters" : {
        "type" : "object",
        "properties" : {
          "command" : {
            "type" : "string"
          },
          "workingDirectory" : {
            "type" : "string"
          }
        },
        "required" : [ "command", "workingDirectory" ]
      }
    }
  } ]
}
2025-10-28 14:10:56,537 INFO  [io.qua.lan.ope.com.OpenAiRestApi$OpenAiClientLogger] (vert.x-eventloop-thread-190) Response:
- status code: 400
- headers: [vary: Origin,Access-Control-Request-Method,Access-Control-Request-Headers, Accept-Encoding], [x-request-id: dc2f77e1-b675-4a18-b6b7-d1d51ad9c60b], [x-dashscope-call-gateway: true], [content-type: application/json], [content-length: 436], [req-cost-time: 253], [req-arrive-time: 1761631856279], [resp-start-time: 1761631856532], [x-envoy-upstream-service-time: 239], [set-cookie: acw_tc=dc2f77e1-b675-4a18-b6b7-d1d51ad9c60b70e39be231237b5bb89e64b05e6738c0;path=/;HttpOnly;Max-Age=1800], [date: Tue, 28 Oct 2025 06:10:56 GMT], [server: istio-envoy]
- body: {"error":{"code":"invalid_parameter_error","param":null,"message":"<400> InternalError.Algo.InvalidParameter: An assistant message with \"tool_calls\" must be followed by tool messages responding to each \"tool_call_id\". The following tool_call_ids did not have response messages: message[6].role","type":"invalid_request_error"},"id":"chatcmpl-dc2f77e1-b675-4a18-b6b7-d1d51ad9c60b","request_id":"dc2f77e1-b675-4a18-b6b7-d1d51ad9c60b"}

The call_2df3eb6ba05644cfbaeadb99 made by assistant was followed by another assistant, but not the expected tool response.
Then I added prompt "Please call the functions serially", but not helping.
I'm using quarkus-langchain4j-openai 1.2.2.

[geoand]

Thanks.

Unfortunately, I don't see any way to help without a sample application that exhibits the behavior you mention