Module 3 Quiz: Agentic Workflows and Coding Assistants

Six questions. Click an option to check it.

---

Question 1

A task has no clear definition of "done" yet. Which mode should you use first?

1. Agent mode - it will figure it out
2. Inline completion
3. Chat, to clarify the intent and write the specification
4. A cloud coding agent attached to the issue

---

Question 2

Where does Claude Code read project-level instructions from?

1. .github/copilot-instructions.md
2. CLAUDE.md at the repository root
3. .cursor/rules/*.mdc
4. package.json

---

Question 3

What does a model actually "see" when it answers?

1. The whole repository
2. Everything on your disk the tool has permission to read
3. Its context window: the prompt, rules, attached files and tool output assembled for that request
4. The git history

---

Question 4

Which of these belongs in a rules file?

1. An explanation of how Python decorators work
2. This sprint's goals
3. "All routes go through app/api/router.py, which applies auth. Do not register routes elsewhere."
4. The database password for the staging environment

---

Question 5

Why is per-user memory a poor substitute for a shared rules file?

1. Memory is slower
2. Memory is invisible to teammates, so the team's conventions end up applied inconsistently
3. Memory cannot store text
4. Memory is only available in Cursor

---

Question 6

An agent is on its third attempt at the same failing test and has started editing a configuration file you did not mention. What should you do?

1. Let it continue; it is close
2. Ask it to try harder
3. Stop the session and restart with a tighter specification
4. Switch to inline completion