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

purpose_agent/purpose_function.py

@@ -105,6 +105,15 @@ Evaluate this state transition. Remember:
 - Score Φ(state_before) and Φ(state_after) on the 0.0–10.0 scale
 - Cite SPECIFIC evidence from the state data
 - Do NOT give credit for intentions — only actual changes
+
+Respond in this format:
+```toml
+phi_before = 0.0
+phi_after = 0.0
+reasoning = "Your step-by-step justification"
+evidence = "Specific state changes you observed"
+confidence = 0.5
+```
 """