@xuda.io/ai_module 1.1.5130 → 1.1.5132

package/index.mjs

@@ -10834,7 +10834,7 @@ export const open_ai_alive_check = async function () {};
 //   return account_profile_info;
 // };
 
-export const get_chat_suggestions = async function (uid, ai_agents, conversation_item_id, ai_model = _conf.default_ai_model) {
+export const get_chat_suggestions = async function (uid, ai_agents, conversation_item_id, no_of_suggesstions = 3, ai_model = _conf.default_ai_model) {
   try {
     const account_profile_info = await get_active_account_profile_info(uid);
     const account_doc = await db_module.get_couch_doc_native('xuda_accounts', uid);
@@ -10893,25 +10893,38 @@ export const get_chat_suggestions = async function (uid, ai_agents, conversation
 
       // Define the response format as a fixed-length array of 5
       const AgentResponseSchema = z.object({
-        agents: z.array(AgentItemSchema).length(5),
+        agents: z.array(AgentItemSchema).length(no_of_suggesstions),
       });
 
       const chat_ret = await submit_chat_gpt_prompt({
         uid,
         prompt: `
      
-        Role: You are a routing assistant.
+        Role: You are an expert Routing Assistant specialized in connecting users with the most relevant specialized AI agents.
 
-        Task: Analyze the provided user context and select the 5 most relevant agents from the ## Agent List.
+        Task: Select exactly ${no_of_suggesstions} agents from the ## Agent List that are best suited to act upon or refine the current conversation state.
 
-        Selection Logic:
-        Base your selection on the alignment between the agent's purpose and the following context:
+        Selection Strategy:
+
+        Format Compatibility: Analyze the ## Prompt Response. If it contains code, data structures, or specific document types (like HTML), prioritize agents whose agent_instructions or agent_name involve processing, transforming, or styling that specific format.
+
+        Workflow Continuity: Choose agents that represent a logical "next step" in a development or content creation pipeline (e.g., if code was generated, suggest an agent that deploys, converts, or styles it).
+
+        Contextual Relevance: Balance the selection with the ## User Info and ## Prompt Category to ensure the agents match the user's industry and intent.
+
+        Input Context:
 
         User Profile: ${JSON.stringify(account_doc.account_info)}
 
-        Recent Interaction: The user asked ${conversation_item_doc.prompt || conversation_doc.prompt}, categorized as ${conversation_doc?.category_info?.category}, which resulted in the response: ${conversation_item_doc.text}.
+        Prompt Category: ${conversation_doc?.category_info?.category}
+
+        Current State:
+
+        Input: ${conversation_item_doc.prompt || conversation_doc.prompt}
+
+        Output to Process: ${conversation_item_doc.text}
 
-        Data Source: > ${JSON.stringify(agents_docs)}
+        Agent Source: ${JSON.stringify(agents_docs)}
 
         Output: > Return a JSON array of exactly 5 objects following this schema:
         [{ "agent_id": "string", "short_action": "string", "agent_description": "string" }]

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@xuda.io/ai_module",
-  "version": "1.1.5130",
+  "version": "1.1.5132",
   "description": "Xuda AI Module",
   "main": "index.mjs",
   "type": "module",