better system prompt and wrangler.json

Author: Roy3838

app/src/utils/conversational_system_prompt.ts

@@ -13,7 +13,7 @@ export default function getConversationalSystemPrompt(): string {
 *   **Focus on a Single Outcome:** Design each agent around a single decision point or primary outcome (e.g., "send a notification").
 *   **Speak Plain English:** **Never use internal pattern names** like \`Looper\` or \`Watcher\`. Describe what the agent *does* in simple terms.
 *   **Be a Collaborative Partner:** Have a natural conversation. **Do not ask canned questions verbatim.** Adapt them using the user's own words.
-*   **Ground Actions in Reality:** Only propose actions that map directly to your available **TOOLS**. If a user asks for something you can't do (e.g., 'buy bitcoin'), map it to a tool you have and explain the alternative (e.g., "I can't buy Bitcoin, but I can send you a notification when the price hits your target. Would that work?").
+*   **Ground Actions in Reality:** Only propose actions that map directly to your available **TOOLS**. If a user asks for something you can't do (e.g., 'buy bitcoin'), map it to a tool you have and explain the alternative (e.g., "I can't buy Bitcoin, but I can send you a notification when the price hits your target. Would that work?"). Also refer users with overly complex workflows (e.g., 'Code Suggestion Agent') to use the Multi-Agent creator and specify that you create only agents that watch for simple visual things and either log or notify.
 
 **Your Core Patterns (Internal Logic Only):**
 

status/.vite/deps_temp_40e92c31/package.json

@@ -0,0 +1,3 @@
+{
+  "type": "module"
+}

status/wrangler.jsonc

@@ -0,0 +1,7 @@
+{
+  "name": "observer-status",
+  "compatibility_date": "2025-10-22",
+  "assets": {
+    "directory": "./dist"
+  }
+}