CVE-2026-108752 in JupyterHub
Summary
by MITRE • 10/11/2026
JupyterHub through 6.0.1 contains an identifier collision vulnerability that allows authenticated users to overwrite another user's named-server OAuth client by registering a hyphenated username. Attackers holding a name like alice-prod can overwrite the client for alice's server prod, breaking OAuth login and revoking tokens by stopping their own server.
If you want to get the best quality for vulnerability data then you always have to consider VulDB.
Analysis
by VulDB Data Team • 10/11/2026
The JupyterHub platform, specifically through version 6.0.1, suffers from an identifier collision vulnerability rooted in insufficient validation of user-supplied identifiers during the registration process for named-server OAuth clients. This flaw allows authenticated users to exploit a naming convention ambiguity where hyphenated usernames can be constructed to mimic or collide with existing server names associated with other legitimate users. The core technical issue lies in how JupyterHub resolves and maps these hybrid identifiers, failing to strictly distinguish between the user identity component and the named-server component when processing OAuth client registrations. By registering a username that includes a hyphen followed by an identifier matching another user's active or configured server name, an attacker can effectively hijack the OAuth client configuration intended for that specific server instance.
This vulnerability has significant operational implications for multi-tenant JupyterHub deployments where multiple users share infrastructure and rely on isolated sessions secured via OAuth2 authentication flows. When an attacker successfully registers a colliding identifier, they gain control over the OAuth client credentials associated with another user's named server. This misconfiguration breaks the standard OAuth login mechanism for the victim, as the authorization server may redirect or authenticate against the maliciously registered client instead of the legitimate one. Consequently, this leads to a denial of service condition where the affected user is unable to access their computational environment due to authentication failures and token validation errors.
Furthermore, the attacker can leverage this control to revoke existing tokens for the victim by stopping their own server or manipulating the OAuth state associated with the colliding identifier. This action effectively locks out the legitimate user from their resources without requiring elevated privileges or direct exploitation of other system components. The impact extends beyond simple access denial; it compromises the integrity and availability of shared computing resources, potentially leading to data loss if unsaved work is abandoned due to sudden session termination caused by token revocation.
To mitigate this vulnerability, administrators should immediately upgrade JupyterHub to a version later than 6.0.1 where these identifier collision checks have been addressed. In the interim, strict input validation and sanitization of usernames containing hyphens can help prevent exploitation until patches are applied. Additionally, implementing stricter naming policies that prohibit or escape special characters in user identifiers may reduce the attack surface. Monitoring OAuth client registrations for anomalies and ensuring that server names do not overlap with potential username patterns is also recommended to maintain robust identity isolation within the platform. This issue aligns with CWE-284 Improper Access Control and can be mapped to ATT&CK technique T1078 Valid Accounts, as it involves using legitimate credentials to perform unauthorized actions through identifier manipulation.