Regarding the issue of using context correlation in Windows AI Foundry ->Phi Silica

How to use context correlation is not well demonstrated in the example, and I am not able to effectively utilize contextual content. Can you provide corresponding support.

I have used ai dev galley. But the examples it provides cannot support contextual content well.

Its example code is to add questions and answers to prompt words.

  private string GetPrompt(IEnumerable<ChatMessage> history)
  {
      if (!history.Any())
      {
          return string.Empty;
      }
  
      string prompt = string.Empty;
  
      var firstMessage = history.FirstOrDefault();
  
      _languageModelContext = firstMessage?.Role == ChatRole.System ?
          _languageModel?.CreateContext(firstMessage.Text, new ContentFilterOptions()) :
          _languageModel?.CreateContext();
  
      for (var i = 0; i < history.Count(); i++)
      {
          var message = history.ElementAt(i);
          if (message.Role == ChatRole.System)
          {
              if (i > 0)
              {
                  throw new ArgumentException("Only first message can be a system message");
              }
          }
          else if (message.Role == ChatRole.User)
          {
              string msgText = message.Text ?? string.Empty;
              prompt += msgText;
          }
          else if (message.Role == ChatRole.Assistant)
          {
              prompt += message.Text;
          }
      }
  
      return prompt;
  }

No repository file, test, or entry point is identified. Start by clarifying the expected context-correlation behavior for Windows AI Foundry and Phi Silica, then review the cited GetPrompt example and determine how contextual content should be represented. Done means documenting or implementing a supported approach with a verifiable example.