Module Check
You have covered ten lessons on the craft of writing prompts that build. You know what a spec is and why it matters. You know to describe outcomes instead of processes, to give context and constraints, to show examples when words fall short, to break big asks into steps, to stay focused with one change at a time, to read and verify what the AI gives back, and to diagnose and fix misunderstandings before they spiral. This review is your chance to confirm what stuck — and to sharpen anything that did not. Work through it deliberately. The flashcards first, then the quizzes, then the capstone.
Key Terms Review
Flashcards — click each card to reveal the answer
Module Quizzes
A friend writes this prompt: 'Make a game.' Which spec component is most critically missing?
You want the AI to build a login form. Which prompt uses better outcome language?
You are building a study app for kindergarteners. Which piece of information is most important to include as context?
Why do examples often communicate intent better than descriptions alone?
You are debugging a complex feature and something broke after your last prompt. You made four changes in that prompt. What is the core problem?
The AI returns a response that does what you asked, but also adds an entire navigation menu you did not request. What should you do?
Writing prompts that build is a craft, and like all crafts it improves with practice. You now have the full vocabulary: spec, outcome language, context, constraints, examples, decomposition, focused requests, acceptance checks, and correction techniques. Every prompt you write from here is an opportunity to apply and refine these skills. The AI is a powerful builder — your job is to be a precise architect.
Capstone — The Complete Prompt Portfolio
- Build a portfolio of three complete, high-quality build prompts — one for each difficulty level below. For each prompt, attach a brief annotation explaining your choices.
- Level 1 — Simple tool:
- Write a build prompt for a word-count tool: the user types or pastes text, and the page shows how many words, sentences, and characters are in it. Annotate: What context did you include? What constraints? Did you use an example?
- Level 2 — Multi-step build:
- Write a decomposed build plan (3-5 prompt steps) for a simple flashcard study app with the ability to add new cards. For each step, name what the user will be able to test when it is done.
- Level 3 — Correction challenge:
- Imagine you sent this prompt: 'Add a dark mode to the page.' The AI made ALL text white and ALL backgrounds black, making images and colored elements look wrong. Write a targeted correction prompt that fixes the problem without rebuilding the whole page. Annotate: what was the misunderstanding, which cause type, and how does your correction address it?
- Review all three prompts using the five-step framework and four evaluation questions from the Workshop lesson. Revise each one at least once before you submit.