THREATOPS
THREAT OPSThreat News › [GHSA] GHSA-x2rj-828p-hx9m (critical) — Xinference vulnerable to remote code execution via unsafe `eval()` in Llama3 tool-call parsing

[GHSA] GHSA-x2rj-828p-hx9m (critical) — Xinference vulnerable to remote code execution via unsafe `eval()` in Llama3 tool-call parsing

medgithub_advisoriesPublished 2026-08-21

GHSA-x2rj-828p-hx9m Severity: critical CVE: CVE-2026-61539

Xinference vulnerable to remote code execution via unsafe `eval()` in Llama3 tool-call parsing

### Summary

Xinference used Python's unsafe `eval()` function when parsing Llama3 tool-call output generated by a large language model. Because the model output can be influenced by attacker-controlled prompts sent to the chat completion API,

MITRE ATT&CK techniques

Indicators of compromise

Original source: https://github.com/advisories/GHSA-x2rj-828p-hx9m