// Security Advisory Triage
Stanza lemmatizer allows RCE via pickle deserialization
GitHub advisories — pip · published 2026-11-11 05:27 UTC
WHAT CHANGEDCVE-2026-115878 published 3 November 2026. The Stanza lemmatizer deserializes untrusted pickle data, enabling remote code execution. Patched in version 1.15.0.
WHY IT MATTERSAn attacker who can control input to the lemmatizer can execute arbitrary code in the Python process. Stanza is a common NLP dependency; exploitation requires no authentication if the lemmatizer processes external input.
WHAT TO DOUpgrade stanza to 1.15.0 or later. If you cannot upgrade immediately, audit all code paths that pass external input to the lemmatizer and block untrusted sources.
Model-reported confidence 95% (not measured accuracy)
Primary source ↗
written by anthropic/claude-sonnet-4.5
Show the reasoning — raw diff, triage verdict, confidence
RAW DIFF
+ghsa_id: GHSA-gh9r-c94j-cmp5 +summary: Stanza: RCE via pickle deserialization in the lemmatizer +severity: high +published_at: 2026-11-03T20:29:16Z +cve_id: CVE-2026-115878 +vulnerabilities[].package.name: stanza +vulnerabilities[].first_patched_version: 1.15.0
HOW IT WAS JUDGED
Triage signals below are model-reported assertions, preserved verbatim; they are not independent arithmetic verification.
| triage verdict | MATERIAL at 0.92, needed 0.5 |
|---|---|
| triage reasoning | RCE via unsafe deserialization in widely-used NLP library affects services running Stanza for text processing. |
| signals | RCE vulnerability · pickle deserialization (known attack surface) · patched version available (1.15.0) · high severity rating · Python package in NLP/ML dependency tree |
| suppression | — |
| triage model | anthropic/claude-haiku-4.5 ($0.00000) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00000) |