Track 3: TOML prompts + PURPOSE_LEARNING.md whitepaper — purpose_agent/actor.py

purpose_agent/actor.py

@@ -73,16 +73,16 @@ ACTOR_STEP_PROMPT = """\
 
 Based on the current state and your goal, decide your next action.
 
-Respond in this exact JSON format:
-```json
-{{
-  "thought": "Your reasoning about the current state and what to do next",
-  "action": {{
-    "name": "action_name",
-    "params": {{"param1": "value1"}}
-  }},
-  "expected_delta": "Specific prediction of what will change in the state"
-}}
+Respond in this format:
+```toml
+thought = "Your reasoning about the current state and what to do next"
+expected_delta = "What you expect to change"
+
+[action]
+name = "action_name"
+
+[action.params]
+param1 = "value1"
 ```
 """