Writing the job around three problems the hire will own in the first six months is much better than a laundry list of technologies. It also makes the paid trial task easier to design, because you already know what "useful" looks like.

We see a similar pattern when founders bring in outside help instead of a full-time hire. The first engagement works best when it is one clearly bounded problem at a fixed price, with acceptance criteria and accounts owned by the company from day one. You learn how the person communicates and delivers, and the cost of finding out is capped.

Did the short paid task end up becoming your main filter, or do you still run a traditional interview alongside it?