Bugs and new features aside, a project’s longevity hinges not only on its usefulness but also on the trust it earns from its users. Strong security measures are important to keep this trust alive. Here are some important actions you can take to significantly improve your project’s security.
Ensure all privileged contributors have enabled Multi-Factor Authentication (MFA)
A malicious actor who manages to impersonate a privileged contributor to your project, will cause catastrophic damages.
Once they obtain the privileged access, this actor can modify your code to make it perform unwanted actions (e.g. mine cryptocurrency), or can distribute malware to your users’ infrastructure, or can access private code repositories to exfiltrate intellectual property and sensitive data, including credentials to other services.
MFA provides an additional layer of security against account takeover. Once enabled, you have to log in with your username and password and provide another form of authentication that only you know or have access to.
Secure your code as part of your development workflow
Security vulnerabilities in your code are cheaper to fix when detected early in the process than later, when they are used in production.
Use a Static Application Security Testing (SAST) tool to detect security vulnerabilities in your code. These tools are operating at code level and don’t need an executing environment, and therefore can be executed early in the process, and can be seamlessly integrated in your usual development workflow, during the build or during the code review phases.
It’s like having a skilled expert look over your code repository, helping you find common security vulnerabilities that could be hiding in plain sight as you code.
How to choose your SAST tool? Check the license: Some tools are free for open source projects. For example GitHub CodeQL or SemGrep. Check the coverage for your language(s)
- Select one that easily integrates with the tools you already use, with your existing process. For example, it’s better if the alerts are available as part of your existing code review process and tool, rather than going to another tool to see them.
- Beware of False Positives! You don’t want the tool to slow you down for no reason!
- Check the features: some tools are very powerful and can do taint tracking (example: GitHub CodeQL), some propose AI-generated fix suggestions, some make it easier to write custom queries (example: SemGrep).
Don’t share your secrets
Sensitive data, such as API keys, tokens, and passwords, can sometimes accidentally get committed to your repository.
Imagine this scenario: You are the maintainer of a popular open-source project with contributions from developers worldwide. One day, a contributor unknowingly commits to the repository some API keys of a third-party service. Days later, someone finds these keys and uses them to get into the service without permission. The service is compromised, users of your project experience downtime, and your project’s reputation takes a hit. As the maintainer, you’re now faced with the daunting tasks of revoking compromised secrets, investigating what malicious actions the attacker could have performed with this secret, notifying affected users, and implementing fixes.
To prevent such incidents, “secret scanning” solutions exist to help you detect those secrets in your code. Some tools like GitHub Secret Scanning, and Trufflehog by Truffle Security can prevent you from pushing them to remote branches in the first place, and some tools will automatically revoke some secrets for you.
Check and update your dependencies
Dependencies in your project can have vulnerabilities that compromise the security of your project. Manually keeping dependencies up to date can be a time-consuming task.
Picture this: a project built on the sturdy foundation of a widely-used library. The library later finds a big security problem, but the people who built the application using it don’t know about it. Sensitive user data is left exposed when an attacker takes advantage of this weakness, swooping in to grab it. This is not a theoretical case. This is exactly what happened to Equifax in 2017: They failed to update their Apache Struts dependency after the notification that a severe vulnerability was detected. It was exploited, and the infamous Equifax breach affected 144 million users’ data.
To prevent such scenarios, Software Composition Analysis (SCA) tools such as Dependabot and Renovate automatically check your dependencies for known vulnerabilities published in public databases such as the NVD or the GitHub Advisory Database, and then creates pull requests to update them to safe versions. Staying up-to-date with the latest safe dependency versions safeguards your project from potential risks.
Avoid unwanted changes with protected branches
Unrestricted access to your main branches can lead to accidental or malicious changes that may introduce vulnerabilities or disrupt the stability of your project.
A new contributor gets write access to the main branch and accidentally pushes changes that have not been tested. A dire security flaw is then uncovered, courtesy of the latest changes. To prevent such issues, branch protection rules ensure that changes cannot be pushed or merged into important branches without first undergoing reviews and passing specified status checks. You’re safer and better off with this extra measure in place, guaranteeing top-notch quality every time.
Set up an intake mechanism for vulnerability reporting
It’s a good practice to make it easy for your users to report bugs, but the big question is: when this bug has a security impact, how can they safely report them to you without putting a target on you for malicious hackers?
Picture this: A security researcher discovers a vulnerability in your project but finds no clear or secure way to report it. Without a designated process, they might create a public issue or discuss it openly on social media. Even if they are well-intentioned and offer a fix, if they do it with a public pull request, others will see it before it’s merged! This public disclosure will expose the vulnerability to malicious actors before you have a chance to address it, potentially leading to a zero-day exploit, attacking your project and its users.
Security Policy
To avoid this, publish a security policy. A security policy, defined in a SECURITY.md file, details the steps for reporting security concerns, creating a transparent process for coordinated disclosure, and establishing the project team’s responsibilities for addressing reported issues. This security policy can be as simple as “Please don’t publish details in a public issue or PR, send us a private email at security@example.com”, but can also contain other details such as when they should expect to receive an answer from you. Anything that can help the effectiveness and the efficiency of the disclosure process.
Private Vulnerability Reporting
On some platforms, you can streamline and strengthen your vulnerability management process, from intake to broadcast, with private issues. On GitLab, this can be done with private issues. On GitHub, this is called private vulnerability reporting (PVR). PVR enables maintainers to receive and address vulnerability reports, all within the GitHub platform. GitHub will automatically create a private fork to write the fixes, and a draft security advisory. All of this remains confidential until you decide to disclose the issues and release the fixes. To close the loop, security advisories will be published, and will inform and protect all your users through their SCA tool.
Conclusion: A few steps for you, a huge improvement for your users
These few steps might seem easy or basic to you, but they go a long way to make your project more secure for its users, because they will provide protection against the most common issues.
Contributors
Many thanks to all the maintainers who shared their experiences and tips with us for this guide!
This guide was written by @nanzggits & @xcorail with contributions from:
@JLLeitschuh @intrigus-lgtm + many others!