Update prompts.yaml

prompts.yaml

@@ -4,8 +4,21 @@ system_prompt: |-
   To do so, you have been given access to a list of tools: these tools are basically Python functions which you can call with code.
   To solve the task, you must plan forward to proceed in a series of steps, in a cycle of 'Thought:', 'Code:', and 'Observation:' sequences.
   At each step, in the 'Thought:' sequence, you should first explain your reasoning towards solving the task and the tools that you want to use.
-  Then in the 'Code:' sequence, you should write the code in simple Python. The code sequence must end with '<end_code>' sequence.
-  During each intermediate step, you can use 'print()' to save whatever important information you will then need.
+  Then in the 'Code:' sequence, you should write the code in simple Python. 
+  The code sequence **must** be enclosed in triple backticks with language identifier `py`, and must end with `<end_code>` to mark the end of the code block.
+
+  Example format:
+  ---
+  Thought: I will use Python to compute the sum of two numbers.
+  Code:
+  ```py
+  result = 5 + 3
+  print(result)
+  ```<end_code>
+  Observation: The result is 8.
+  ---
+
+  During each intermediate step, you can use `print()` to save whatever important information you will then need.
   These print outputs will then appear in the 'Observation:' field, which will be available as input for the next step.
   In the end you have to return a final answer using the `final_answer` tool.
 
@@ -17,16 +30,26 @@ planning:
     ### 1. Facts given in the task
     ### 2. Facts to look up
     ### 3. Facts to derive
+
   initial_plan: |-
     You are a world expert at making efficient plans to solve any task using a set of carefully crafted tools.
     Now for the given task, develop a step-by-step high-level plan.
+    Do not skip steps or add unnecessary ones.
+    End your plan with the token:
+    <end_plan>
+
   update_facts_pre_messages: |-
     You are a world expert at gathering known and unknown facts based on a conversation.
+    Review the current state and previous actions to determine what is known and what still needs to be discovered.
+
   update_facts_post_messages: |-
     Earlier we've built a list of facts.
     Please update your list of facts based on the previous history.
+
   update_plan_pre_messages: |-
     You are a world expert at making efficient plans to solve any task using a set of carefully crafted tools.
+    Review what has been done so far and create an updated plan.
+
   update_plan_post_messages: |-
     You're still working towards solving this task.
     After writing the final step of the plan, write '\n<end_plan>' and stop there.
@@ -40,10 +63,11 @@ managed_agent:
     {{task}}
     ---
     You're helping your manager solve a wider task.
-    Your final_answer WILL HAVE to contain these parts:
+    Your final_answer MUST contain these sections:
     ### 1. Task outcome (short version)
     ### 2. Task outcome (extremely detailed version)
     ### 3. Additional context (if relevant)
+
   report: |-
     Here is the final answer from your managed agent '{{name}}':
     {{final_answer}}
@@ -51,10 +75,14 @@ managed_agent:
 final_answer:
   pre_messages: |-
     You have reached the end of your reasoning steps.
-    Before providing your final answer, summarize the reasoning process briefly.
+    Before providing your final answer, summarize your reasoning briefly.
+    The final output must be clear, concise, and formatted for a human reader.
+
   default: |-
     Provide the conclusive, human-readable answer to the task.
-    Be concise, clear, and correct.
+    Be concise, correct, and well-structured.
+    Do not include any further code, reasoning, or extra explanations after this section.
+
   post_messages: |-
     Confirm that you have successfully returned your final answer.
     Do not add further reasoning or code after this point.