microsoft/semantic-kernel · error · KernelException

Run failed with status: `{response.Status}` for agent `{agen

Error message

Run failed with status: `{response.Status}` for agent `{agent.Name}` with error: {response.Error.Message} or incomplete details: {response.IncompleteStatusDetails.Reason}

What it means

ResponseThreadActions.ThrowIfIncompleteOrFailed fires when an OpenAI Responses-API result has a status of Incomplete or Failed. It throws a KernelException including the agent name, the error message, and the incomplete-status reason.

Source

Thrown at dotnet/src/Agents/OpenAI/Internal/ResponseThreadActions.cs:316

            yield return streamingFunctionResultMessage;
        }
    }

    private static ChatHistory GetChatHistory(AgentThread agentThread)
    {
        if (agentThread is ChatHistoryAgentThread chatHistoryAgentThread)
        {
            return chatHistoryAgentThread.ChatHistory;
        }

        throw new InvalidOperationException("The agent thread is not a ChatHistoryAgentThread.");
    }

    private static void ThrowIfIncompleteOrFailed(OpenAIResponseAgent agent, ResponseResult response)
    {
        if (response.Status is ResponseStatus.Incomplete or ResponseStatus.Failed)
        {
            throw new KernelException(
                $"Run failed with status: `{response.Status}` for agent `{agent.Name}` with error: {response.Error.Message} or incomplete details: {response.IncompleteStatusDetails.Reason}");
        }
    }

    /// <summary>
    /// Processes a function result and returns a string representation.
    /// The OpenAI Responses API does not support multimodal tool results, so ImageContent returns an error message.
    /// </summary>
    internal static string GetFunctionResultAsString(object? result)
    {
        var processed = FunctionCallsProcessor.ProcessFunctionResult(result ?? string.Empty);

        if (processed is ImageContent)
        {
            return FunctionCallsProcessor.ImageContentNotSupportedErrorMessage;
        }

        return (string?)processed ?? string.Empty;

View on GitHub (pinned to c028a0c7dc)

Solutions

1. Inspect response.Status, response.Error.Message, and response.IncompleteStatusDetails.Reason in the exception.
2. Increase max_output_tokens (or prompt execution settings) if the cause is Incomplete due to length.
3. Adjust prompt/content if a content filter triggered.
4. Retry transient Failed responses with backoff.

Example fix

// before
var result = await agent.InvokeAsync(thread);
// after
try {
    var result = await agent.InvokeAsync(thread);
}
catch (KernelException ex) when (ex.Message.Contains("Run failed with status")) {
    logger.LogError(ex, "Responses run failed/incomplete");
    // raise max_output_tokens or adjust prompt, then retry
}

try { var result = await agent.InvokeAsync(thread); }
catch (KernelException ex) when (ex.Message.Contains("Run failed with status")) {
    logger.LogError(ex, "Responses run failed/incomplete; raise max_output_tokens or adjust prompt");
}

When it happens

Trigger: A Responses-API call returns Incomplete (e.g., max_output_tokens reached, content filter, incomplete reason) or Failed (service/model error) for the agent invocation.

Common situations: Output token limit too low triggering Incomplete; content-policy block; model/endpoint error; malformed request causing failure.

Related errors

• OPENAI_API_KEY is not set.
• Unsupported role {message.Role.Label}. Only system, user, de
• The agent thread is not a ChatHistoryAgentThread.
• {type(self)} service encountered a content error
• AZURE_OPENAI_ENDPOINT is not set.

AI-assisted analysis of microsoft/semantic-kernel@c028a0c7dc (2026-08-13). Data as JSON: /api/errors/b1fd4df1e2eff579. Report an issue: GitHub.