Developing an application isn't a cake walk, although it can seem like one if all we're talking about is getting a program up and running. Perhaps the biggest chore with developing an app is securing it tight, analyzing the code until you arrive at Eye Strain City. Security is king. It's also time-consuming.
MIT knows this reality all too well, so it's decided to do something to help ease the pain. The result is a system called CodePhage, one that takes advantage of other applications to help bolster the security in its own. I wouldn't blame you if you thought that such mechanics were impossible, but indeed they are.
Rather than just analyze source code from other projects and assume snippets would be worth implementing into one's own app, CodePhage analyzes the behavior of certain apps to see exactly what it is they're doing, such as which security checks have been put into place.
What's beneficial about this design is that because the source code isn't required, CodePhage can reap the rewards of any application - the language it was coded in simply doesn't matter. That said, CodePhage can ultimately import freely available code from projects that matches the goals of its confirmed checks. Such code imports wouldn't be taken lightly; any imported code would have to be a "common" solution, one that many apps share identical implementations of. MIT is not about to snoop around proprietary code and snatch what it can, but instead would focus on truly open sourced projects.
There are many more specifics about the entire process than I've tackled here, so if you want to learn more, I'd highly recommend hitting up the MIT article.