I like Jimmy's answer, but feel like it could be extended upon further. Not knowing how much experience the person has, turn the "simple project" into something that can show basic and extended ability (as well as demonstrate how eager they are to learn technologies when they're not familiar).
What I mean by that is (again, borrowing @Jimmy's example) you may ask for one thing, but provide a "bonus" version of it shortly thereafter. For instance:
- Creating functional pages is one thing, while asking for a uniform/polished styling is extra credit.
- Being able to create login form, but adding validation to the form may be the extra credit.
- Mandating a specific user/password is used for authentication, while having them establish a database and a register form (using a database) would be the extra credit.
So on and so forth. This allows you to gain insight into a few metrics:
- Basic demonstration of ability
Were they able to tackle some of the bonus tasks, and, if so, how many/which were they comfortable implementing?
- Task prioritization
Were all required tasks done and were working on bonuses, or were they doing each one serially and didn't get some of the final tasks done.
- What questions were asked (if any)
Did your mentor developer get approached on some of the topics they were unfamiliar with.
- Enthusiasm/self learning
Did s/he do the bare minimum or try to do more (did they do the CSS themselves or use a Twitter Bootstrap)? You may learn what they are really efficient at just by the process they used.
I might even place an item at the bottom (separated from the list, but in the closest following paragraph) about optionally using source control. Something that describes you'd like them (if comfortable) to use the source control and check-in between tasks. From there, if/how it was used gives you a little more insight:
- Not Used
- Mentor not approached
Probably didn't even read the paragraph (keep this in mind for task that require micro and macro understandings).
- Mentor approached
Read paragraph, but not sure what it is.
- Used
- One big check-in
Last minute thing, they work-top to bottom without deviation.
- Individual check-ins per item
Read the entire requirement first (bonus: you have insight into how they work.)