NLTK hat eine Path Traversal Vulnerability (AFO) - Arbiträres Überschreiben von Dateien

The core of the vulnerability lies in the NLTK downloader's insufficient validation of the subdir and id attributes within remote XML index files. An attacker controlling a malicious XML index server can craft responses containing path traversal sequences (e.g., ../). This allows them to bypass intended directory restrictions and perform several malicious actions. Specifically, attackers can create directories at arbitrary locations on the file system, create arbitrary files, and, critically, overwrite existing system files. Overwriting files like /etc/passwd or ~/.ssh/authorized_keys could grant attackers unauthorized access and control over the affected system. The potential for arbitrary file overwrite significantly elevates the risk associated with this vulnerability.

Systems running NLTK versions 3.9.2 and earlier are at risk, particularly those where the NLTK downloader is exposed to untrusted XML index servers. Development environments and automated build pipelines that utilize NLTK are also potential targets.

• python / nltk:

import os
import hashlib

def check_nltk_version():
    import nltk
    version = nltk.version.version
    if version <= '3.9.2':
        print(f"NLTK version {version} is vulnerable to CVE-2026-33236.")
    else:
        print(f"NLTK version {version} is not vulnerable.")

check_nltk_version()

• generic web: Monitor access logs for requests to NLTK download endpoints containing path traversal sequences (e.g., ../). • generic web: Check for unexpected files or directories created in system directories.

1. 2026-03-19Disclosure

   disclosure

1. Reserviert2026-03-18
2. Veröffentlicht2026-03-19
3. Geändert2026-03-25
4. EPSS aktualisiert2026-03-28

The primary mitigation for CVE-2026-33236 is to upgrade to a patched version of NLTK. Consult the NLTK project's release notes for the latest version. If upgrading is not immediately feasible, consider implementing strict input validation on the subdir and id attributes within the NLTK downloader. This could involve whitelisting allowed characters or enforcing maximum path lengths. Additionally, restrict access to the NLTK downloader to trusted sources and networks. Carefully review any remote XML index files before processing them to identify and reject malicious content. After upgrading, confirm the fix by attempting to download a resource using a known malicious path traversal sequence and verifying that the download fails with an appropriate error.