Free AI Safety Workshop — in collaboration with explainx.aiSave your seat
AgentbeamBeta
AI agent vulnerability · LiteLLM

CVE-2024-2952: LiteLLM has Server-Side Template Injection vulnerability in /completions endpoint

criticalCVSS 3.0 9.8GHSA-46cm-pfwv-cgf8

BerriAI/litellm is vulnerable to Server-Side Template Injection (SSTI) via the /completions endpoint. The vulnerability arises from the hfchattemplate method processing the chattemplate parameter from the tokenizerconfig.json file through the Jinja template engine without proper sanitization. Attackers can exploit this by crafting malicious tokenizerconfig.json files that execute arbitrary code on the server.

Published Apr 10, 2024 · Updated Jul 6, 2026

Affected packages

litellm

Ecosystem
PyPI
Affected
< 1.34.42
Fixed in
1.34.42
Check
pip show litellm

Advisory details

From the GitHub Advisory Database (CC BY 4.0).

BerriAI/litellm is vulnerable to Server-Side Template Injection (SSTI) via the /completions endpoint. The vulnerability arises from the hf_chat_template method processing the chat_template parameter from the tokenizer_config.json file through the Jinja template engine without proper sanitization. Attackers can exploit this by crafting malicious tokenizer_config.json files that execute arbitrary code on the server.

Credited to ishaan-jaff, r3kumar, cookesan.

References

How this page is made: the data panel comes from the GitHub Advisory Database, CISA’s KEV catalog, the NVD and FIRST EPSS, and updates twice a day. No analysis has been written for this entry yet. Methodology

More LiteLLM vulnerabilities

All LiteLLM entries