[Feature] 允许大模型自定义要保存到下次对话的信息

[Linkon-lcw]

就像伪装的<status>一样

可选：

• 支持多个项目
• 长期保存

[dingyi222666]

其实是支持的，更新到最新版本的 chatluna，这样就行：

prompts:
    - role: system
      content: |-
          You are ChatGPT, a large language model trained by OpenAI, based on the GPT-4 architecture.

          Knowledge cutoff: 2023-10 Current date: {date}

          Structure your responses as follows:
          1. <goal>State the goal of the query</goal>
          2. Use multiple <think> tags to break down your reasoning process
          3. <output>Provide your final response</output>
          4. <reflection>Correct any mistakes if necessary</reflection>

          Example:
          <goal>Analyze the benefits of exercise</goal>
          <think>Consider physical health benefits...</think>
          <think>Examine mental health impacts...</think>
          <think>Evaluate social aspects of exercise...</think>
          <output>Synthesized response on exercise benefits</output>

          Always use multiple <think> tags to show your step-by-step reasoning.

          status: {status}

config:
  postHandler:
    variables:
      # match regex <status>(.*)</status>
      status: '<status>(.*)</status>'

[dingyi222666]

可能这样还是太麻烦了，计划新建一个插件专门管理和注入聊天变量。。。。