CVE-2026-126295

An NLTK parser could open model files beyond its configured data roots

NLTK project · Natural Language Toolkit BllipParser model loading

First documented disclosure · Review checkpoint

Summary

The parser passed unchecked model paths to its native dependency. The parser could open sensitive host files or block a worker on a special file.

Severity and exploitation evidence

High · 7.0/10 · CVSS 4.0. Severity source.

Exploitation not established: The reviewed advisory does not report malicious exploitation.

Exploitation evidence source

Technical severity is not a measure of observed attacks or of risk to your deployment. Attempts, demonstrations, and successful compromise retain the qualifications in the source.

How the weakness works

The parser passed unchecked model paths to its native dependency.

Conditions required

An application allowed an untrusted party to choose BLLIP model or reranker paths.

Potential impact

The parser could open sensitive host files or block a worker on a special file.

Affected versions and configurations

3.10.0 through 3.10.3.

Fix and mitigation status

3.10.4.

Practical next steps

Upgrade NLTK to 3.10.4 or later. Limit model locations to trusted directories.

Operating lesson

Apply data-root restrictions before passing paths to native model parsers.

Evidence and limits

The CVE record was published on 2026-11-13.

Sources