authentik CVE-2026-49443

Two concrete prerequisites are required: (1) the attacker must hold an authentik account with permission to change a source connection (a source-edit/management capability, typically scoped to administrators or delegated source managers, not granted to ordinary users by default), and (2) the attacker must also possess a valid account in at least one of the external authentication sources configured in the target authentik instance (e.g., an OAuth/SAML/LDAP identity the IdP federates). Exploitation is limited by the requirement for an existing low-privileged but source-edit-capable authentik account (PR:L), which prevents fully unauthenticated exploitation, and by the deployment having external sources configured - instances using only built-in password authentication with no federated sources are not exposed via this vector.

The CVSS 3.1 vector AV:N/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H signals a network-reachable, low-complexity attack that requires only low privileges and no user interaction, with high impact across the CIA triad - consistent with a full account-takeover primitive. PR:L is the key gating factor: an attacker needs both an authentik account with permission to edit a source connection (typically an administrative or delegated-management role) and a valid identity in one of the configured sources, which limits opportunistic exploitation but is highly dangerous in multi-tenant or delegated-admin deployments. EPSS data was not provided in the input and the CVE is not in CISA KEV, so real-world exploitation pressure is currently unmeasured; the absence of a public PoC further suggests this is not yet under mass exploitation, but the impact (any-account takeover on an IdP) means a single successful exploit can cascade into every downstream application authentik protects, warranting prompt patching.

An attacker holding a low-privileged authentik account that has been granted the ability to edit source connections (for example, a delegated tenant admin or a compromised helpdesk role) and possessing a valid identity in one of the configured external sources (e.g., a personal Google or GitHub account) modifies the source binding so that their federated identity maps to a target user - such as the authentik superuser or a privileged downstream application owner - then completes a normal federated login flow and is authenticated as that victim. No public exploit code has been identified at time of analysis, but the attack requires no special tooling beyond the authentik admin UI and a legitimate account in the abused source.