oss-sec mailing list archives
Re: xdg-open bypassing SameSite=Strict
From: Lucas Holt <luke () foolishgames com>
Date: Tue, 24 Jun 2025 16:48:53 -0400
On 6/24/25 4:22 PM, Gabriel Corona wrote:
I would think that all browsers should implement the safe behavior or URL handler registrations and allow the user (or enterprise) to adjust the policy within settings. This would limit the issue for the vast majority of users, but allow folks to turn on the old behavior until applications can be fixed. I suspect this could break some auth flows that rely on handlers right now within many apps. (game launchers, enterprise tools, etc)As was said by Solar Designer, if a "safe" version is needed, it should probably be the default when going through URI scheme registrations. This is because, as you said, this kind of issue lies in the interaction between several components (URI sources, URI sinks and URI go-betweens such as xdg-open) and it would certainly be possible to find a way to bypass the behavior otherwise.
I don't think the right place to handle this is xdg-open. Essentially, browsers are trusting unsafe input as is.
-- Lucas Holt Luke () FoolishGames com ________________________________________________________ MidnightBSD.org (Free OS) JustJournal.com (Free blogging)
Current thread:
- xdg-open bypassing SameSite=Strict grape mingijung (Jun 23)
- Re: xdg-open bypassing SameSite=Strict Solar Designer (Jun 23)
- Re: xdg-open bypassing SameSite=Strict grape mingijung (Jun 24)
- Re: xdg-open bypassing SameSite=Strict Simon McVittie (Jun 24)
- Re: xdg-open bypassing SameSite=Strict Anton Luka Šijanec (Jun 24)
- Re: xdg-open bypassing SameSite=Strict Gabriel Corona (Jun 24)
- Re: xdg-open bypassing SameSite=Strict Lucas Holt (Jun 24)
- Re: xdg-open bypassing SameSite=Strict Solar Designer (Jun 23)
