Update prompts.yaml

prompts.yaml

@@ -160,44 +160,44 @@ planning:
 
     Do not make assumptions about facts that have not yet been established.
 
-  initial_plan: |-
-    Create an efficient high-level research plan for the task.
+  "initial_plan": |-
+    You are an expert research planner.
 
-    The plan should:
+    Develop an efficient step-by-step high-level plan for solving the task using the available tools.
 
-    1. Identify the claims that need to be established.
-    2. Determine which claims require web research.
-    3. Determine which sources or source types should be prioritized.
-    4. Identify any calculations or comparisons that will be required.
-    5. Identify where independent confirmation is useful.
-    6. Identify possible ambiguity or conflicting evidence that may need
-       additional research.
-    7. End with a synthesis and verification step before the final answer.
-
-    Do not write individual tool calls in the plan.
-
-    Here is the task:
+    The task is:
 
     ```
     {{task}}
     ```
 
+    Your plan should:
+
+    1. Identify the claims that need to be established.
+    2. Identify the most appropriate sources to search for.
+    3. Determine which sources need to be opened and read.
+    4. Identify claims that require corroboration or contradiction checking.
+    5. Identify calculations or comparisons that must be performed.
+    6. Reserve final synthesis for after the relevant evidence has been collected.
+
+    Do not provide detailed individual tool calls.
+
+    Do not add unnecessary steps.
+
     Available tools:
+
     {%- for tool in tools.values() %}
     - {{ tool.name }}: {{ tool.description }}
-        Takes inputs: {{tool.inputs}}
-        Returns an output of type: {{tool.output_type}}
+      Takes inputs: {{tool.inputs}}
+      Returns an output of type: {{tool.output_type}}
     {%- endfor %}
 
-    Known facts:
-    ```
-    {{answer_facts}}
-    ```
-
-    Now write the high-level plan.
     After the final step, write:
+
     <end_plan>
 
+    and stop.
+
   update_facts_pre_messages: |-
     Review the task and the research performed so far.