Every day brings one kata: a function stub and a hidden stack of tests. You start with the first test only.
Write just enough code to turn it green, then press Run tests. Each run counts toward your score.
When everything shown is green, reveal the next test. Revealing is free, and if your code already handles it, it lights up green straight away.
Clear every test in as few runs as you can. Par is one run per test; a general solution can beat it.
Once the last test passes, runs stop counting. Refactor as much as you like, then share your grid.
Stuck? Anything you console.log shows up in the console with a tab per test, the line that printed it, and repeats folded into a count. Click an object to expand it. After a run, the selected test's values appear beside the lines that logged them. console.table, count, time, group and assert work too. Type a call like fizzBuzz(15) into the console prompt to try your own inputs; each call is tallied as a probe on your result.
Shortcuts: Ctrl ↵ runs the tests, Ctrl / toggles comments, Tab and Shift Tab indent, Shift Alt F formats. Suggestions pop up as you type; Ctrl Space brings them back, and Tab or Enter picks one. Click a line number in an error to jump to it.