Update api/server.py

api/server.py

@@ -143,8 +143,8 @@ async def step(action: Action):
         return StepResponse(
             observation = environment._build_observation(),
             reward      = Reward(
-                score     = 0.001,
-                breakdown = {"validation_error": 0.001},
+                score     = 0.0,
+                breakdown = {"validation_error": 0.0},
                 feedback  = f"Malformed action: {str(e)}"
             ),
             done = False,
@@ -196,7 +196,7 @@ async def grader(request: GraderRequest):
     try:
         if request.action is None:
             return GraderResponse(
-                score     = 0.001,
+                score     = 0.0,
                 feedback  = "No action provided for grading.",
                 breakdown = {"error": "null_action"}
             )
@@ -215,7 +215,7 @@ async def grader(request: GraderRequest):
             score = round(
                 (perf_improvement * 0.60) + (step_efficiency * 0.20) + 0.10, 4
             )
-            score = max(0.001, min(0.999, score))
+            score = max(0.0, min(1.0, score))
 
             return GraderResponse(
                 score    = score,
@@ -233,12 +233,12 @@ async def grader(request: GraderRequest):
 
         # Round 1 grading
         score, breakdown, feedback = grade(request.action, request.task_id)
-        score = max(0.001, min(0.999, score))
+        score = max(0.0, min(1.0, score))
         return GraderResponse(score=score, feedback=feedback, breakdown=breakdown)
 
     except Exception as e:
         return GraderResponse(
-            score     = 0.001,
+            score     = 0.0,
             feedback  = f"Grader error: {str(e)}",
             breakdown = {"error": str(e)}
         )
@@ -331,4 +331,4 @@ async def root():
         "tasks_count": 30,
         "max_steps":   50,
         "themes":      ["Long-Horizon Planning", "World Modeling", "Self-Improvement", "Wildcard"],
-    }
+    }