Set minimum shaped reward to 0.001 (strict >0)

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

backend/src/polypharmacy_env/rewards.py

@@ -93,4 +93,4 @@ def compute_shaped_reward(
 
     # finish_review terminal bonus is added by the caller after grading
 
-    return max(0.0, reward)
+    return max(0.001, reward)