NLTK dependency vulnerability

[okal]

My team's dependabot scanner picked up on a vulnerability (Critical level) from the nltk dependency. It's been fixed in 3.9.3 (draft PR here #922).

[devin-ai-integration]

Hi Okal (@okal), thank you for reporting this issue!

I've confirmed the vulnerability you're referencing: CVE-2025-14009 — a critical (CVSS 10.0) Zip Slip vulnerability in the NLTK downloader component (nltk/downloader.py). The _unzip_iter function uses zipfile.extractall() without path validation, which could allow an attacker to craft a malicious zip package that executes arbitrary code when downloaded and extracted by NLTK. The fix is available in nltk 3.9.3.

There is also a second CVE on versions prior to 3.9.3: CVE-2026-0848 (Unsafe Dependency Resolution in the StanfordSegmenter module).

I can see your PR #922 to bump nltk to 3.9.3 — it looks like the right fix. That PR is currently closed with some CI failures; the Airbyte team will evaluate the best path to get this merged.

This issue is already being tracked internally and has been escalated to the engineering team:

https://github.com/airbytehq/oncall/issues/11795

---

Need more help? Join Airbyte Community Slack for peer support, or if you're a Cloud customer, open a support ticket referencing this URL.