CVE-2026-129221

Training labels could alter the schema of an NLTK Weka dataset

NLTK project · Natural Language Toolkit · Weka classifier

First documented disclosure · Review checkpoint

Summary

A contributor to a training set could alter the structure of the file NLTK handed to the Weka classifier.

Severity and exploitation evidence

Medium · 5.4/10 · CVSS 3.1. Severity source.

Exploitation not established: The NLTK advisory describes the vulnerability and its preconditions. The reviewed sources identify no malicious use.

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 Weka adapter wrote class labels into an ARFF schema without encoding delimiters.

Conditions required

Untrusted contributors supply labels to a Weka training pipeline.

Potential impact

A contributor could change the schema or data supplied to training, or cause training to fail.

Affected versions and configurations

<=3.10.3

Fix and mitigation status

3.10.4.

Practical next steps

Upgrade to NLTK 3.10.4 or a later fixed release.

Operating lesson

Validate external training labels and encode them for the dataset format.

Evidence and limits

The issue-specific patch was public by 2026-08-19. The fixed release shipped October 21. The November 15 maintainer advisory supplies the identifier and affected versions. This is a historical addition.

Sources