Loading course title...

Loading assessment title...

Connecting to LMS... Progress: in progress

Assessment

1. Which description best captures vibe coding as taught in this course? A. Letting an AI assistant make all implementation decisions once a goal is mentioned. B. Writing code manually while avoiding AI suggestions until the final review. C. Using conversation to direct software changes while the developer reviews, tests, and approves the result. D. Replacing design, testing, and code review with a sequence of longer prompts.

2. A generated feature works in one quick demonstration. Why is that not enough evidence for production use? A. The demo covers only one path and may miss edge cases, security issues, or violated requirements. B. The feature should be rejected because AI-generated code can never be useful. C. A successful demo proves the code is correct but not fast enough. D. The assistant should be asked for a longer explanation instead of running tests.

3. What belongs in a useful task description for an AI coding assistant? A. Only the desired feature name, so the assistant has maximum creative freedom. B. A long list of unrelated files, regardless of whether they affect the change. C. A request to edit every layer at once so the assistant has full context. D. The user, goal, inputs, outputs, environment, constraints, and observable success criteria.

4. Which context is usually most helpful when adding a feature to an existing repository? A. A general article about the programming language with no project-specific details. B. The relevant source files, tests, interfaces, errors, dependency versions, and project conventions. C. Every file in the repository without explaining the requested change. D. Only the newest file in the project, because older files are never authoritative.

5. Why should large AI-assisted changes be decomposed into smaller steps? A. Small steps make failures easier to isolate and each patch easier to inspect, test, and recover. B. Small steps prevent the assistant from producing any incorrect code. C. Small steps remove the need for version control. D. Small steps are mainly useful when the developer does not plan to review the result.

6. During code review, what should the developer inspect? A. Only whether the code compiles successfully. B. Only the assistant's explanation of the change. C. Patch scope, architecture, validation, error handling, boundaries, resource use, and dependencies. D. Only the files that have the largest number of changed lines.

7. Which information best supports evidence-driven debugging with an AI assistant? A. A short message saying the feature is still broken. B. The exact command, relevant error output, expected behavior, observed behavior, reproducible input, and recent changes. C. A request for the assistant to rewrite the entire feature using a different style. D. A screenshot alone, even when the failure happens in a command-line tool.

8. Why must suggested dependencies be verified before installation? A. Every dependency suggested by an assistant is maintained by the same organization. B. Verification is needed only for paid software libraries. C. A dependency is safe if its name sounds familiar. D. A package may be unnecessary, outdated, incompatible, poorly maintained, wrongly licensed, insecure, or nonexistent.

9. What separates a useful prototype from production-ready software? A. A prototype becomes production-ready as soon as it demonstrates the main idea. B. Production readiness mostly means the code was generated quickly enough. C. Production software has been hardened for real users, operations, security, testing, documentation, and future maintenance. D. A prototype is production-ready if it avoids comments and documentation.

10. Which sequence best represents the disciplined vibe coding loop? A. Define intent, provide context, plan when needed, make a small change, inspect, validate, debug from evidence, review risks, document, then approve or revise. B. Prompt broadly, accept the first running version, skip tests, and return later if users report problems. C. Ask for a full rewrite, install suggested packages, and deploy once the assistant says the work is finished. D. Focus on prompt wording while leaving scope, testing, security, and maintainability for another team.