Write A Requirements Card Before Prompting

End-of-lesson check

15 questions · take it digitally for instant feedback at tendril.neural-forge.io/learn/quiz/end-vibecoder-requirements-card

1. What five elements make up a requirements card for AI prompting?

   1. User, job, data, edge case, and success check
   2. Prompt, response, model, temperature, and tokens
   3. Frontend, backend, database, API, and testing
   4. Name, price, date, author, and version

2. Why does the lesson recommend naming the job before naming the tool?

   1. AI models process job descriptions more accurately than tool names
   2. To ensure the tool serves the actual need rather than forcing a solution to fit a predetermined tool
   3. Because tools are more expensive than defining jobs
   4. To make the prompt shorter and more efficient

3. What does 'smallest useful scope' mean when writing a requirements card?

   1. Minimizing the number of test cases to speed up validation
   2. Writing the shortest possible prompt in terms of character count
   3. Building the minimum feature set that provides real value to users
   4. Using the cheapest AI model available for cost savings

4. Why should you run AI output as a user rather than as a fan of the tool?

   1. To evaluate whether the result actually solves the user's problem, not just whether the tool worked
   2. Because user feedback is automatically more valuable than developer feedback
   3. To ensure the AI gets credit for good work
   4. To avoid discovering bugs that would be embarrassing to reveal

5. What three things should you inspect before sharing an AI-generated solution?

   1. Model version, temperature, and token limit
   2. Color scheme, font choice, and layout
   3. Price, speed, and accuracy metrics
   4. Diff, data access, and failure path

6. What is scope creep in the context of AI-assisted development?

   1. When testing takes significantly longer than the development phase
   2. When the AI model file size grows too large to deploy
   3. When users request features that the AI cannot implement
   4. When a project gradually expands beyond its original boundaries, making it unmanageable

7. In the context of Supabase, what does RLS stand for and why does it matter?

   1. Rapid Learning System, which helps the AI improve over time
   2. Resource Loading System, which manages server performance
   3. Row Level Security, which controls who can access what data in the database
   4. Remote Logic Service, which handles API requests

8. Why is a rollback plan essential when deploying AI-generated code?

   1. Because AI models always produce incorrect code on the first try
   2. It allows you to revert to a working state if the AI output causes problems
   3. To reduce hosting costs by shutting down failed deployments
   4. To satisfy compliance requirements that mandate backup systems

9. Which question about data should a requirements card explicitly answer?

   1. Who trained the underlying AI model?
   2. How much does the API call cost?
   3. What is the current date and time?
   4. What data should the app or agent never expose?

10. What is an edge case in requirements planning?

    1. The first and most important test case in a suite
    2. The most common way users interact with the system
    3. An unusual condition or scenario that requires special handling beyond the normal flow
    4. The simplest possible example of a feature working

11. What is the 'success check' component of a requirements card?

    1. A test that proves the change works as intended
    2. A method to make the AI produce output faster
    3. A metric for measuring developer productivity
    4. A celebration when the AI completes a task

12. What is a prompt brief?

    1. A list of all possible prompts ever written for a project
    2. A concise document that tells the AI exactly what job to accomplish and what constraints exist
    3. A guide to installing and configuring AI tools
    4. A long legal document about AI usage rights

13. Why is 'hidden public data' identified as a common breakpoint in AI app development?

    1. It makes the application run faster by caching data publicly
    2. It reduces hosting costs by eliminating private databases
    3. It improves user experience by making more data readily available
    4. It exposes information that should be private but is accidentally accessible to anyone

14. What does 'brittle fix loops' mean in the context of AI-generated code?

    1. Efficient code that is easy to maintain long-term
    2. Robust solutions that handle errors gracefully
    3. Fragile cycles where each fix creates new problems, leading to endless iteration
    4. Fast debugging processes that resolve issues in one pass

15. What should the 'rollback path' element of a requirements card specify?

    1. The specific steps to revert changes if the AI output is incorrect or causes issues
    2. The order in which features should be implemented
    3. The path the AI follows to generate its response
    4. The user journey through the application interface