新拉取的main代码配置ollama使用qwen2.5vl上传图片推理报错

新拉取的main代码配置ollama使用qwen2.5vl上传图片推理报错
单agent（自主规划模式），输入文字没有问题，输入图片就报错

1、模型在文字推理下没有问题

2、在图片推理下报错

3、后台错误日志

No file, test, or entry point is named. Reproduce the single-agent autonomous-planning case with Ollama and qwen2.5vl, comparing the successful text-only path with image inference and inspecting the backend error shown in the issue screenshots. Done means image inference completes without the reported error.

Written by the indexing model from the issue text.