Procoder

1. What is the central principle of Spec-Driven Development (SDD)?

• A. Start implementing code, then write spec afterwards for documentation

• B. Write tests first, ignoring spec until after code is functional

• C. Create a detailed specification of what the system must do before any code or test is written

• D. Let GenAI generate prompt templates, then derive spec from AI outputs

✔ Correct Answer: Create a detailed specification of what the system must do before any code or test is written

2. In an AI First workflow (using GenAI / LLMs) without SDD + TDD, what issue is most likely to occur?

• A. The feature will have thorough test coverage from the start

• B. Rapid prototyping but with mismatches between user expectations and behavior

• C. Very slow development due to writing too many specs upfront

• D. Perfect alignment between stakeholders and implementation

✔ Correct Answer: Rapid prototyping but with mismatches between user expectations and behavior

3. Which sequence correctly represents combining SDD + TDD in a GenAI workflow?

• A. Write spec → generate code via AI → write tests → run tests

• B. Generate code → write spec → write tests → run tests

• C. Write spec → derive tests → write tests → prompt AI to generate code → run tests

• D. Derive tests → write code manually → adjust spec afterwards

✔ Correct Answer: Write spec → derive tests → write tests → prompt AI to generate code → run tests

4. Which non-functional requirement is most important to include in the spec when doing AI First Development?

• A. The developer’s preferred editor themes

• B. Readability, maintainability, performance of generated code

• C. Only which libraries the AI should use

• D. The specific variable names to be used throughout

✔ Correct Answer: Readability, maintainability, performance of generated code

5. When should tests be written in a workflow?

• A. After AI has generated the full code, to debug issues

• B. Based on spec, before requesting code generation from AI

• C. Based on spec, after requesting code generation from AI

• D. Only after the feature is deployed and used

✔ Correct Answer: Based on spec, before requesting code generation from AI

6. Which scenario is least appropriate for applying full SDD + TDD workflows?

• A. Building a mission-critical backend service with many stakeholders

• B. Quickly prototyping UI look & feel where behavior is fluid

• C. Developing internal tools required for long-term maintenance

• D. Features requiring compliance, security, and reliability

✔ Correct Answer: Quickly prototyping UI look & feel where behavior is fluid

7. What is the role of human review in AI-generated code when using spec + test workflows?

• A. To only check if code compiles

• B. To check code against spec and tests, handle edge cases, style, security issues

• C. To adjust prompts before generating spec

• D. To skip tests if code looks correct

✔ Correct Answer: To check code against spec and tests, handle edge cases, style, security issues

8. What is a key danger if GenAI code generation is used without spec or strong test framework?

• A. Everything works perfectly, no corrections needed

• B. Hidden security risks, missing authorization/input validation issues

• C. AI always produces slower, less optimized code

• D. Complete spec coverage automatically

✔ Correct Answer: Hidden security risks, missing authorization/input validation issues

9. What does “single source of truth” mean in SDD?

• A. The spec, tests, documentation are all separate and evolve independently

• B. Spec drives the work; code, tests, documentation refer back to it so everything is aligned

• C. AI agent is the single source of generation in the workflow

• D. Only code matters; spec is optional

✔ Correct Answer: Spec drives the work; code, tests, documentation refer back to it so everything is aligned

10. How does SDD w/ TDD affect feedback loops in development?

• A. Feedback mostly happens after deployment in production

• B. Feedback through spec → tests → prompt → code allows early detection of ambiguity or defects

• C. No feedback is needed if spec is well written once

• D. Feedback through spec → code→ prompt → test allows early detection of ambiguity or defects

✔ Correct Answer: Feedback through spec → tests → prompt → code allows early detection of ambiguity or defects

11. What kind of spec writing improves the usefulness of AI First Development?

• A. Very vague statements like “Make UI nice and fast”

• B. Clear, example-rich specs with expected outputs, variants, error messages

• C. Specs that only specify technologies (e.g. “use React, Node.js”) without behavior

• D. Forget testability; only worry about visual style

✔ Correct Answer: Clear, example-rich specs with expected outputs, variants, error messages

12. Which choice best describes how non-functional requirements (NFRs) show up in SDD + GenAI workflows?

• A. They are ignored until after code is built

• B. Spec includes NFRs (performance, security, readability) so prompts/code/test cover them

• C. NFRs only in tests after user complains

• D. NFRs only documented but never enforced

✔ Correct Answer: Spec includes NFRs (performance, security, readability) so prompts/code/test cover them

13. Considering “edge case testing”, which test case is best derived from a spec for a “parse integer from user input” feature?

• A. Input “123” → returns integer 123

• B. Input “-5” → returns negative integer or error based on spec; input “abc” → error; input blank or null → error

• C. Input huge number that fits within bounds only; skip invalid tests

• D. Only test inputs that behave normally

✔ Correct Answer: Input “-5” → returns negative integer or error based on spec; input “abc” → error; input blank or null → error

14. Why is spec writing especially important when using GenAI tools?

• A. Because AI always infers spec correctly from minimal instruction

• B. To provide explicit instructions so AI agents produce reliable, predictable code and to reduce security/risk issues

• C. To allow AI to generate spec itself without human oversight

• D. Because spec replaces all tests

✔ Correct Answer: To provide explicit instructions so AI agents produce reliable, predictable code and to reduce security/risk issues

15. How should spec changes be handled when a behavior diverges during development with GenAI + TDD?

• A. Ignore spec; adapt code to what AI produced

• B. Modify spec or tests so they reflect the correct intended behavior; update prompts; then regenerate or adjust code/tests

• C. Delete tests; rely just on QA feedback

• D. Only document changes without modifying spec or tests

✔ Correct Answer: Modify spec or tests so they reflect the correct intended behavior; update prompts; then regenerate or adjust code/tests