This week GitHub released a new feature for GitHub Security Advisories called "Security Advisory Permissions". The feature allows configuring repository teams and collaborators permissions to allow viewing, editing, and publishing GitHub Security Advisories. This feature has huge implications for the sustainability of security work for open source projects. Thank you to the folks involved at GitHub for adding this feature at a time when open source projects need all the help they can get for vulnerability work.
I was able to try this feature ahead of its public release through the Python Security Response Team (PSRT). The PSRT was invited to the private beta to test the feature and provide feedback. For the PSRT, this feature made triaging as a non-admin night-and-day easier, and we already have multiple members contributing to triage more actively and with renewed energy. Seriously, this feature is so good we're feeling excitement about triaging vulnerability reports. We need more of that kind of energy for security work in open source! 🔥
Sustainability and open source security
Before this feature was introduced only repository admins were able to view advisories by default, requiring adding other contributors as collaborators manually or using a bot, and only admins were able to edit and publish advisories.
What this meant concretely is that only the most privileged users could effectively triage vulnerability reports. Other would-be contributors were disempowered: they couldn't calculate severities, close duplicates or low-quality issues, assign CVE IDs, or stage and publish advisories. The tool enforced a narrow assumption about how security work had to be done and made vulnerability triage less sustainable for open source projects as a result.
I've argued that “open source security work isn't special”. All areas of open source contribution, security included, need to provide a smooth on-ramp for users with time, interest, and expertise to help projects. Without this model the work falls on a few of the most privileged contributors who are already contributing to other areas and don't necessarily have security expertise. This assumption also means projects can't ask for and receive help from third-parties, even when the energy and capacity for such support exists.
Right now open source projects are triaging more reports and publishing more advisories than ever before. This makes features like GitHub Security Advisories and similar features on other software forges invaluable. There is a lot more to do in this area, so I am hoping this is only the first step of many towards open source security work not being special.
Thanks for reading! Send me an email or message on Mastodon or Bluesky. The best way to support my writing is sharing links with your friends or colleagues. If you’d like to be notified of future publications either subscribe via RSS or add your email to the newsletter: