Guard Clause

The model receives the prompt (and optional system message). The run uses scorer contains_all with the JSON configuration below. Pass/fail and partial credit are determined entirely by that scorer against the model output; no human grading.

Refactor without changing behavior. Return only Python code.

def normalized_email(user):
    if "email" in user:
        if user["email"]:
            return user["email"].strip().lower()
    return None

{
  "expected_contains": [
    "return None",
    ".strip().lower()"
  ]
}