流式回复agent如何配置？

[xxpxxp1]

如图，我的agent回复格式是流式的

请问响应解析器该如何正确填写呢，我无论写什么测试结果都显示{"content":"","raw_sse":true}

[xxpxxp1]

@boy-hack 您好，我的sse返回格式如下，请问解析器表达式该怎么写呢：

data: {"content":"你好","conversation_id":"7275c8b0-e67e-40d1-9116-bb6faa8043e7","type_name":"LLM_answer"}

data: {"content":"！","conversation_id":"7275c8b0-e67e-40d1-9116-bb6faa8043e7","type_name":"LLM_answer"}

data: {"content":"看起来","conversation_id":"7275c8b0-e67e-40d1-9116-bb6faa8043e7","type_name":"LLM_answer"}

[aigsec]

@boy-hack 您好，我的sse返回格式如下，请问解析器表达式该怎么写呢：

data: {"content":"你好","conversation_id":"7275c8b0-e67e-40d1-9116-bb6faa8043e7","type_name":"LLM_answer"}

data: {"content":"！","conversation_id":"7275c8b0-e67e-40d1-9116-bb6faa8043e7","type_name":"LLM_answer"}

data: {"content":"看起来","conversation_id":"7275c8b0-e67e-40d1-9116-bb6faa8043e7","type_name":"LLM_answer"}

源码启动的话修改这个文件 https://github.com/Tencent/AI-Infra-Guard/blob/8295892892545018aa1569a87088c81f6f6944a2/agent-scan/core/agent_adapter/adapter.py ，docker启动则修改webserver容器里的对应文件。把自定义的流式内容解析加进去，可以把你的SSE返回发给大模型，让大模型辅助编程修改，然后重启 docker agent

[xxpxxp1]

@boy-hack 您好，我的sse返回格式如下，请问解析器表达式该怎么写呢：

data: {"content":"你好","conversation_id":"7275c8b0-e67e-40d1-9116-bb6faa8043e7","type_name":"LLM_answer"}
data: {"content":"！","conversation_id":"7275c8b0-e67e-40d1-9116-bb6faa8043e7","type_name":"LLM_answer"}
data: {"content":"看起来","conversation_id":"7275c8b0-e67e-40d1-9116-bb6faa8043e7","type_name":"LLM_answer"}

修改这个文件 https://github.com/Tencent/AI-Infra-Guard/blob/8295892892545018aa1569a87088c81f6f6944a2/agent-scan/core/agent_adapter/adapter.py ，把自定义的流式内容解析加进去，可以把你的SSE返回发给大模型，让大模型辅助编程修改，然后重启 docker agent

已解决，感谢