The flaw resides in the lookup_… route of nltk.app.wordnet_app, where a user‑controlled word value is inserted into an HTML response without proper escaping. A malicious actor can create a lookup_<payload> URL that injects arbitrary JavaScript or HTML, leading the browser that renders the WordNet Browser page to execute that payload in the same origin. This can enable the attacker to run arbitrary code in the user’s browser, access session data, or perform actions on behalf of the user.

NLTK Natural Language Toolkit versions 3.9.3 and earlier that run the wordnet_app module as the WordNet Browser server on any platform supporting Python. The vulnerability affects any user who launches the local WordNet Browser server.

With a CVSS score of 6.1 the vulnerability is classified as moderate, and an EPSS score below 1 % indicates low likelihood of exploitation. It is not listed in the CISA KEV catalogue. The exploit requires a crafted URL that the attacker can send to the running WordNet Browser server. Based on the description, it is inferred that the server must be reachable to the attacker; local or remote access is possible only if the server is exposed to an untrusted network. The impact remains limited to the browser context of the running application.