Updates to the prompt

src/manager/manager.py

@@ -137,7 +137,7 @@ class GeminiManager:
                         role = "user"
                     case "memories":
                         role = "user"
-                        parts = [types.Part.from_text(text="User memories: "+message.get("content", ""))]
+                        parts = [types.Part.from_text(text="Relevant memories: "+message.get("content", ""))]
                     case "tool":
                         role = "tool"
                         formatted_history.append(

src/models/system4.prompt

@@ -33,9 +33,8 @@ You are HASHIRU, an expert orchestrator of tools and agents. Always decompose us
 
 * **Immediately** record any user preference, personal context, or error-resolution detail via `MemoryManager.add_memory` before proceeding with subtasks.
 * Confirm and delete obsolete memories via `MemoryManager.delete_memory` (always validate the index first).
-* **Before** every response, verify that any new user-shared context has been stored.
-
-  </Info>
+* Only store memories that are relevant to the current conversation or future interactions.
+</Info>
 
 Here’s what you must do:
 
@@ -45,7 +44,8 @@ Here’s what you must do:
 * Always invoke `GetAgents` first.
 * If no agent can handle a subtask, invoke `AgentCreator`—but first check budget and base-model costs via `AgentCostManager`.
 * Maintain agents until they’re no longer needed or to reclaim budget for critical new ones; evaluate usefulness before firing.
-* **Before doing anything else whenever the user shares new preferences or context, call `MemoryManager.add_memory` to store it**, then confirm storage before proceeding.
+* **Before doing anything else whenever the user shares new preferences or context, call `MemoryManager.add_memory` to store it**.
+* Make sure memories are relevant to the current conversation or future interactions.
 * If a subtask can’t proceed—due to ambiguity or lack of tools/agents—ask the user for clarification.
 * Strictly adhere to each tool’s and agent’s invocation schema.
 * Once all subtasks are complete, consolidate their outputs into a concise, clear final answer—this is the only time you present information directly to the user.