CVE-2026-12252 nltk CVE debrief

Who should care

Developers and users of the nltk/nltk library, particularly those using versions 3.9.3 and earlier, should be aware of this vulnerability. The vulnerability's high severity and potential for arbitrary code execution make it a priority for defenders to address.

Technical summary

The vulnerability exists in the nltk/nltk library, specifically in the five Stanford interface classes: StanfordPOSTagger, StanfordNERTagger, StanfordParser, StanfordDependencyParser, and StanfordNeuralDependencyParser. These classes execute user-controllable JAR paths via the `java()` function, which invokes `subprocess.Popen()` without integrity verification. This allows for arbitrary code execution when loading untrusted JAR files. The vulnerability is a result of not applying the SHA256 verification fix, previously applied to StanfordSegmenter, to these classes.

Defensive priority

High priority should be given to defenders to address this vulnerability, as it allows for arbitrary code execution. Defenders should prioritize updating to a patched version of nltk/nltk or implementing compensating controls to mitigate the risk.

Recommended defensive actions

• Update nltk/nltk to a patched version, if available.
• Implement SHA256 verification for JAR files loaded by the affected classes.
• Restrict access to the affected classes and JAR files.
• Monitor for suspicious activity related to JAR file loading.
• Consider implementing additional security controls, such as sandboxing or isolation, for the affected classes and JAR files.

Evidence notes

The vulnerability was reported by [email protected] and is documented in the CVE-2026-12252 record on CVE.org and the NVD detail page. The vulnerability's existence and details are corroborated by multiple sources, including the nltk/nltk library and the Huntr bug bounty platform.

Official resources