Om sesjonen
OAuth consents have become a daily chore for most of us. We all know about consent phishing, but what about the case where the OAuth client is legitimate, comes from a trusted provider, and has already been granted access to your resources?
Developers, admins, and security professionals normally focus on what we can control: the hardening of our Identity Providers and Resource Servers. However, the moment a user or admin clicks "Accept" on a third-party integration, the attack surface is effectively increased to include that third party, a black box of code, infrastructure, and governance which we have zero control over. In this session, we will dive into how this trust can be abused to compromise apps and organizations and how to limit it for your own resources.
We will demonstrate some examples from Entra ID and Google that basically delegate top-tier admin roles to third parties. This is not an edge case as we see it in virtually all organizations we work with. As for application-specific vulnerabilities, we will show practical examples from bug bounty findings at Google, demonstrating how I could steal all your spreadsheets if you consented to a first-party Google OAuth client.