Create a new authentication type which allows to have a short session by opening a link sent by email.
That's improve the security because emails can already be used as mean of password recovery and this authentication avoid to users to set not convenient to use passwords.
Description
Status | Subtype | Assigned | Task | ||
---|---|---|---|---|---|
Open | None | T336 Add login capabilities through external services | |||
Open | None | T668 E-mail authentification |
Event Timeline
This is especially interesting as it's passwordless and so allow to avoid any credential in the db.
A recent post about that: Quincy Larson, https://www.freecodecamp.org/news/360-million-reasons-to-destroy-all-passwords-9a100b2b5001/#:~:text=Remember%20Myspace%3F,password%20combinations%20on%20other%20websites., Free Code Camp blog.
Development moratoire
Per T1771, we're currently considering implementing Keycloak as a reference identity management and SSO login product.
This product exposes a LDAP, OIDC (OpenID Connect) and SAML capabilities to authenticate users and applications. It seems to solve our main problems.
From there, it's not clear what we do with Auth Grove:
- Scenario A. We drop it, and as users we directly interact with Keycloak. Development is discontinued.
- Scenario B. Auth Grove is morphed into a front-end to use Keycloak: we expose current information, and interact with Keycloak API (through a generic set of classes to allow to switch to another solution) to set credentials and metadata.
- Scenario C. We use both Keycloak AND Auth Grove. We don't make integrate with Keycloak at all, to stay independent and not vendor-locked.
While T1771 evaluation is ongoing, a moratoire covers any development activities related to Auth Grove, with the obvious exception of security issues.
This moratoire cover fully or partly this task.