Ship Assistent disk persist, park LLM, and memory UI cleanup.

Split the extension into partials, persist chats on the data volume, park/warm the chat model around Generate, and drop dual raw/persona dump paths.

Co-authored-by: Cursor <cursoragent@cursor.com>

AssistentVram.cs

@@ -57,7 +57,7 @@ public partial class SwarmAssistentExtension
     }
 
     /// <summary>Single-token chat so the model is resident again by the time the user types.</summary>
-    public async Task<JObject> AssistentWarmLlm(Session session, string baseUrl, string model, string persona = null)
+    public async Task<JObject> AssistentWarmLlm(Session session, string baseUrl, string model)
     {
         string root = NormalizeBaseUrl(baseUrl);
         string name = (model ?? "").Trim();