[insoluble]

> It can be hard to tell how much someone actually contributed just by looking at a CV.

While this is true, it can also be hard to tell how much of a person's open-source code was actually written by that person. With the abundance of code available today, it would be relatively trivial for someone to take random code fragments and functions from all around and to assemble them together into a conglomerate. All the person would have to do is basically change the variable and function names, in addition to the syntactic style, to make it all appear as coming from the same person. Sure, this would at least show an awareness of syntax and style, but it certainly would not show the level of programming skill or even necessarily of problem-solving ability. On the other hand, if the person's repo is popular, then there is at least an indication of function and utility.

[atroyn]

There is also having pulls accepted on open source, e.g. fixing bugs in various libraries and so on - one doesn't need to start a whole OSS project from scratch to contribute.