View Source, Bindiff for Vuln Analysis, Bypass with GitHub Actions, & NIST DevSecOps – ASW #170

Mike Shema

1. Missouri governor faces backlash and ridicule for threatening reporter who discovered exposed teacher SSNs
   Journalism is an important tool against abuse of political power and holding power to account. In this case, a journalist discovered that their state's education web site exposed payment data and SSNs for its teachers. The governor's response -- a belligerent misstatement of the flaw and threats of legal action -- hearken back to early days of appsec before coordinated vulnerability disclosure was a term of art and it was never sure whether an org's security team (if one existed at all) or legal team would respond to a vuln report. Orgs don't need to dive right into bug bounty programs to handle vuln disclosure, but they should take the steps towards a solution like that. One resource for this kind of maturity model is at https://www.lutasecurity.com/vcmm
2. Student finds zero-days in Exterity devices while rick-rolling school district
   In our other vuln disclosure article of the week, a student discovered flaws in his school's IoT devices -- which just happened to be network-connected TVs, which just happened to be the perfect vector for the infamous rick-roll. The outcome this time around was fortunately more collaborative. Even so, there are still lessons in this story about coordinated disclosure and needing productive responses from vendors to patch vulns. Read the student's write up at https://whitehoodhacker.net/posts/2021-10-04-the-big-rick And, of course, refresh your 80s vibe with the video at https://youtu.be/dQw4w9WgXcQ
3. Bindiff and POC for the IOMFB vulnerability, iOS 15.0.2
   While this is a rather technical write-up, it's still accessible for readers who aren't experts in iOS internals or reverse engineering. It's a nice example of using binary diffs to analyze a patch in order to understand the vuln it fixes. Then, given that understanding, develop a working exploit. Even though the flaw itself seems simple -- an integer overflow in a size calculation -- the thought process to analyze and exploit it are more complex. This article does a great job in walking through those steps.
4. Bypassing required reviews using GitHub Actions
   Here's another article that fits into a recurring theme of security issues stemming from a SaaS vendor's features that can impact your data even if you're not using those features. Consequently, it's also a good example of vendor responsibility in the shared security responsibility model developed by cloud service providers. A positive aspect about this vuln is that it demonstrated what a successful vulnerability disclosure program looks like. A researcher discovered the flaw, shared it with GitHub via their bug bounty platform, and GitHub resolved the issue and rewarded the researcher. All in all the modern model we like to see in such situations.
5. The ‘Leak’ of Warzone’s New Anti-Cheat System Was Actually Part of the Plan
   Infosec in general has many adversarial threat models. Gaming is an interesting area of appsec that has to focus on far more than just secure software practices. Many games also need to counter cheating and abuse in order to keep their games enjoyable and playable for a large population. It's a good exercise in evaluating threat models that go beyond simple injection-style attacks and start to consider ways of countering unintended use or abuse of "business logic" within a game.
6. Implementation of DevSecOps for a Microservices-based Application with Service Mesh
   NIST is collecting feedback on an upcoming guidance for securing the CI/CD pipelines for microservices and containers. It's a bit dense and, by necessity, covers the topic at a high level. However, it can be a useful resource for ensuring your appsec program covers CI/CD well and provides the tools and processes needed to maintain security from writing code to deploying it. It also has useful abstractions in addressing infrastructure as code, policy as code, and observability as code. In other words, it's not shying away from expressing security properties within various areas of a system so that they can be automatically configured, analyzed, and reported on. That kind of automation is a far better future than yet another hardening checklist.
7. SAML explained: How this open standard enables single sign on
   This article doesn't have as neat of a thematic tie-in with the rest of the ones we cover this week, but it still falls into the category of education. We've covered SAML and OAuth a few times in the past, so it makes sense to share an article that provides an easy introduction to their main concepts.