You can build a working app by prompting. Nobody has taught you the ten minutes of checking that stands between a working app and a £3,400 bill, a leaked customer list, or a weekend of 10p cakes.
These are drills, not lessons. Each one drops you into a real-feeling project where an AI assistant has already done something confident and wrong. You prompt it, you inspect what it actually did, and you fix it.
Nothing is graded on what you type. Every check reads the state of the project — is the key really out of the code, do the tests really catch a bug — because that is how reality grades you too.