We will be moving our Topicbox to a self hosted discourse. Our topicbox is very open (in the sense that most subscribers are approved, but initial postings are moderated), but we also don’t want problems of spam, hackers, etc on the new discourse. On discourse we intend to allow authentication via Google, github, and a few trusted sources.
I see differing opinions on this subject from years past regarding the risks of enabling auth via discourse ID. So I am wondering, what is the current wisdom about allowing login using discourse ID? Do the benefits outweigh the potential downsides/risks or as bad as as the google search results suggest? Are the risks of discourse ID any worse that allowing login via the broad GH and google sources?
Those listed risks are mostly what people call “allucinations”.
Reputation Mismatch: Trust levels (TL0–TL4) earned on other forums may not align with your community’s standards, potentially granting external users posting or moderation privileges they haven’t earned locally.
Discourse ID doesn’t do that
Data Dependency: You rely on external servers or networks for identity verification. If the remote login provider experiences downtime or changes its API rules, your users may lose access.
Same is true for every social login
Chained Vulnerabilities: Your site’s security posture becomes tied to the weakest participating site or the central identity provider. If an external partner site is compromised, bad actors may leverage those accounts to access your instance.
Discourse ID is built on top of Discourse, which is same software your community will run.
Account Takeover Propagation: If a user uses a weak password or suffers a credential leak on a remote participating site, that compromised state can flow directly into your forum authentication flow.
This makes no sense.
Use Discourse ID if you want to get many social logins with zero effort. Don’t use it if you want to control exactly what methods users can use to login into your site.
