CVE-2026-129223

A deep WordNet relation tree could crash an NLTK processing job

NLTK project · Natural Language Toolkit · WordNet relation trees

First documented disclosure · Review checkpoint

Summary

An NLP service could lose a worker when it turned a deeply nested lexical relation into a tree.

Severity and exploitation evidence

High · 7.5/10 · CVSS 3.1. Severity source.

Exploitation not established: The reviewed project publications document the weakness. They report no known 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 WordNet tree and acyclic-tree functions used recursive helpers without checking the maximum traversal depth.

Conditions required

A caller can submit a custom WordNet corpus or influence relation data that an application traverses.

Potential impact

Computing a tree could exhaust the Python call stack and stop the job.

Affected versions and configurations

Through 3.10.3.

Fix and mitigation status

3.10.4.

Practical next steps

Upgrade to 3.10.4 or a later fixed release and keep relation-depth limits appropriate to the corpus.

Operating lesson

Check depth even when a relation graph has no cycles.

Evidence and limits

The public August 26 commit bounded recursive walks. An October follow-up strengthened the limit. The November 15 advisory supplies the affected range and the CVSS rating. This is a historical addition.

Sources