Assessment Generation

MCQ Pattern

- stem: "What does OAuth2 refresh token do?"
  options:
    a: "Authenticates user directly"  # distractor
    b: "Obtains new access token"     # correct
    c: "Stores user password"         # distractor
    d: "Encrypts API calls"           # distractor
  answer: b
  bloom: understand
  objective: "Explain OAuth2 token lifecycle"

Coding Question Pattern

- title: "Implement rate limiter"
  prompt: "Write a function that limits requests to N per minute"
  starter: |
    def rate_limit(max_requests: int):
        pass
  tests:
    - input: "5 requests in 1 sec"
      expect: "3 blocked"
  rubric:
    - "Correct logic (50%)"
    - "Edge cases (25%)"
    - "Code quality (25%)"
  bloom: apply

Distractor Guidelines

• Plausible but wrong
• Common misconceptions
• Avoid "all of above" / "none of above"
• Similar length to correct answer

Question Distribution

Gotchas

• 4 options optimal for MCQ
• Randomize answer positions
• Test the test: have SME review