One thing that was a little challenging is that the 'expected' and 'got' values are reversed in the test cases. So it took me awhile to realize that what was 'expected' was actually what my code was returning, and what was 'instead got' was actually the correct answer. Might be good to fix this.

Fun one. Inspired me to investigate other interpreters.

Still some typos in the description (deferred vs. deffered).

It's unclear what the expected support for multiple callbacks is, and the error message: testing multiple callbacks - resolve & reject should call the success - could be more descriptive.

Agree that leaving out the loop skeleton, making the number of arrays arbitrary, and setting up some test cases would all improve this.