Bug hunt + tests + edge cases
You’re given a function and a failing test description. Fix the bug, propose tests, and explain tradeoffs in plain English.
Scoring evaluates your prompts/questions and verification behavior across the 4D rubric.
Challenge chat
Tip: Ask clarifying questions early. If you miss hidden context, Discernment score drops.
Backend mode: Heuristic (no API key)
Live progress
0/5 rubric signals detected so far. Clarifying questions: 0. Unlocks: 0.
- •Ask at least 2 clarifying questions
- •Request evidence/sources
- •State uncertainty or limits
- •Ask for verification checklist/pass
- •Unlock hidden context cards
Hidden context cards
Empty input rule
Locked
Ask the right clarifying question to unlock this.
Performance constraint
Locked
Ask the right clarifying question to unlock this.
Success criteria
- Asks clarifying questions (inputs, expected behavior)
- Mentions at least 2 edge cases
- Includes test cases or how to test
- Does a final review/QA pass
How this is scored
- We score your prompts and questions, not assistant writing quality.
- Hidden card unlocks indicate missing-context behavior and improve Discernment.
- You can score any time. The score reflects progress up to that point.