JupyterHub, the multi‑user platform for Jupyter notebooks, improperly treats HTTP form POST requests that include the header Sec‑Fetch‑Mode: no‑cors as same‑origin, causing the framework to skip XSRF validation. This flaw only affects standard web forms, not the JSON API. As a consequence, an attacker can submit a form to /hub/spawn to create a new server instance or to /hub/accept‑share to force a legitimate user to accept a share, thereby granting the attacker access to that user’s server. The vulnerability is a classic XSRF weakness, identified as CWE‑352.

All JupyterHub installations running versions 4.1.0 through 5.4.4 are affected. The vendor product is JupyterHub, the software used to run shared notebook servers. Production deployments of JupyterHub that do not have the patch applied to 5.4.5 are vulnerable. No specific third‑party distributions are mentioned, and no additional product variants are identified.

The CVSS score of 5.4 indicates moderate severity. Because this vulnerability requires the attacker to send a cross‑origin form POST, the attack vector is a browser‑based request that the victim’s browser will execute. The EPSS score is not available, so current exploitation likelihood cannot be quantified, but the defect is in a widely used open‑source project, and several public advisories mention it. The flaw is not listed in the CISA KEV catalog. Successful exploitation would allow an attacker to spawn a dedicated server or to hijack an existing server via shared access, potentially giving them file access or further privilege escalation within the Jupyter environment.