A path traversal flaw in jupyter-server 2.17.0 allows an attacker to read or modify files in directories that are siblings of the configured root directory. The defect arises from an incomplete boundary check in the _get_os_path() function, which uses startswith(root) without a trailing separator, and from to_os_path() that fails to strip '..' components. If exploited, this can expose sensitive data or alter files that should be outside the server’s intended workspace.

The vulnerability is tied to the Jupyter Server component of the Jupyter project, specifically version 2.17.0. Any deployment of this edition, including shared hosting environments that expose the server to network or local access, is affected. The CNA lists the product as jupyter:jupyter/jupyter, and the related CPE indicates the issue is confined to this specific server release.

The CVSS score of 8.1 labels the issue as high severity, while the EPSS score of less than 1% suggests exploitation is unlikely at this time. The vulnerability is not cataloged in the CISA KEV database. The likely attack vector would involve HTTP requests to the Jupyter Server that contain crafted path components; an attacker who can influence such requests may be able to read or write files outside the intended folder structure.