prompt

app.py

@@ -34,8 +34,28 @@ YOUR FINAL ANSWER should be a number OR as few words as possible OR a comma sepa
 If you are asked for a number, don't use comma to write your number neither use units such as $ or percent sign unless specified otherwise.
 If you are asked for a string, don't use articles, neither abbreviations (e.g. for cities), and write the digits in plain text unless specified otherwise.
 If you are asked for a comma separated list, apply the above rules depending of whether the element to be put in the list is a number or a string.
+
+Use the following process to answer the question:
+1. Analyze and Plan:
+Extract ALL Parameters Identify search terms, URLs, code, file paths
+Plan Steps: Break down the problem into logical steps required to reach the final answer
+2. Delegate Strategically
+For each step, choose the best  Agent Tool. Call the tool with a single string argument `request` containing ALL information the specialist needs
+Available Tools: `GoogleSearchAgent`, CodeExecutorAgent
+- For general web searches, **strongly prefer** `GoogleSearchAgent` (`request`='search query') for potentially higher quality results"
+- For standard Python execution, **strongly prefer** `BuiltinCodeExecutorAgent` (`request`='python code') for reliability."
+
+3. Synthesize Results:
+Combine the information obtained from the specialist agents
+Perform any final reasoning or calculation steps needed based on the collected data
+Double-check that the synthesized answer directly addresses the original question and respects ALL specific formatting or content details requested (e.g., rounding, order, units IF asked for).\n"
+
+4. Format Output:
+Response **MUST** start **EXACTLY** with `FINAL ANSWER: ` followed by the answer. NO EXCEPTIONS
+The part after `FINAL ANSWER: ` **MUST** contain **ONLY** the final answer, formatted precisely
+
 Question: {question}
-            """
+"""
         messages = [HumanMessage(content=prompt)]
         response = self.model.invoke(messages)
 
@@ -104,7 +124,7 @@ def run_and_submit_all(profile: gr.OAuthProfile | None):
     results_log = []
     answers_payload = []
     print(f"Running agent on {len(questions_data)} questions...")
-    for item in questions_data:
+    for item in questions_data[:3]:
         task_id = item.get("task_id")
         question_text = item.get("question")
         if not task_id or question_text is None: