Open source software has the potential to be very secure. Unlike proprietary code that can only be accessed directly by its own developers, anyone can vet open source projects to spot flaws and bugs. In practice, though, being open source is no panacea. Now, code repository GitHub is rolling out new tools for its GitHub Advanced Security suite that will make it easier to root out vulnerabilities in the open source projects managed on its platform.
Open source code present a few security challenges. In practice there aren’t always enough people with the right expertise looking at it. And open source projects are generally ad hoc; they don’t necessarily have a clear process in place for people to submit vulnerabilities, or the resources available for someone to patch them. Even if you surmount those hurdles, you may not know who’s actually using your open source code and needs a patch.
“A lot of what we talk about is there’s a vulnerability, what’s the workflow for that vulnerability, now it gets addressed,” says Jamie Cool, vice president of product for security for Microsoft-owned GitHub. “But the nirvana is you don’t introduce the vulnerability to begin with. You stop it from ever showing up. It really seems like this is a problem we should be able to help developers not introduce again and again, but by and large we haven’t succeeded at that as a software industry yet.”
In September, GitHub acquired the code scanning tool Semmle as part of a plan to help the GitHub community catch common security flaws automatically. Advanced Security includes this service, calling out which line of code contains a potential vulnerability, why it might be exploitable, and how to fix it. In addition to this automatic scanning, Semmle’s technology can also be used manually by security researchers. GitHub’s goal is to use Advanced Security as both a warning system for developers and a built-in framework for bug hunters to find and report additional issues.
GitHub Advanced Security also includes tools that scan user “repositories,” essentially the folder where they store their development projects, for secret data like passwords and private keys that shouldn’t be exposed and accessible. GitHub works with a number of partners, including Amazon Web Services and Alibaba, to understand the characteristics of their authentication tokens and spot them automatically. The feature has already been available to public repositories for a couple of years, but today GitHub is also adding support to scan private repositories as well. GitHub says that eight percent of active public repositories had a secret exposed in them during the last month alone.
With these new tools, GitHub is working to address security issues at a vast scale. Though not all open source projects rely on GitHub, the majority do, and the platform is as much a social network for the community as a development tool. By offering features like Advanced Security, GitHub can create an environment where more projects in the diverse landscape of open source have access to the same types of tools large companies build to improve and safeguard their proprietary code.
“The truth is for most maintainers they become maintainers by accident,” says GitHub CEO Nat Friedman. “They make something, it becomes widely used and then suddenly they’re in this position of responsibility with regards to computer security—maybe for banks, for governments. They may not have a background in security and yet we have to make sure that the code they publish is secure. So the challenge is to make it automatic and make it natural.”
Though catching more security flaws across GitHub projects is crucial, the interconnected nature of software today still poses security challenges. Rather than writing every function and component from scratch, virtually every software product contains a mix of proprietary code and open source components. Your fitness tracker and smartphone, not to mention your car, all contain open source elements from numerous developer projects in addition to the hardware and software created by the name brand.
Reporting vulnerabilities and getting the right patches to the right places are still prominent problems, because of these interdependencies. In November, GitHub launched a toolkit called Security Lab to help the community keep track of bugs more easily and automate more of the patching process.
While GitHub is in a position to make a major impact on how the open source community handles security, Chris Wysopal, chief technology officer of the software auditing firm Veracode, points out that the progress GitHub is making doesn’t let the rest of the industry off the hook.
“The thing about GitHub is it’s inherently open, so something to improve the landscape of open source doesn’t have to be done by GitHub,” Wysopal says. “There’s nothing stopping a third-party from scanning all of the GitHub repos, looking for vulnerabilities, and sending information to those project maintainers.”
That would take a lot of resources. GitHub itself says it costs millions of dollars to provide the free vulnerability scanning and analysis tools in Advanced Security. The company hopes, though, that its own investment can serve as a model for why it pays to prioritize security in open source.
More Great WIRED Stories
- The devastating decline of a brilliant young coder
- Zoom not cutting it for you? Try exploring a virtual world
- Anti-quarantine protests aren’t about Covid-19
- How to cover your tracks every time you go online
- 26 hours on a Saharan freight train
- ? AI uncovers a potential Covid-19 treatment. Plus: Get the latest AI news
- ? Things not sounding right? Check out our favorite wireless headphones, soundbars, and Bluetooth speakers
