CVE-2026-61539
- Source
- https://cve.org/CVERecord?id=CVE-2026-61539
- Import Source
- https://storage.googleapis.com/cve-osv-conversion/osv-output/CVE-2026-61539.json
- JSON Data
- https://api.osv.dev/v1/vulns/CVE-2026-61539
- Aliases
- Published
- 2026-08-21T20:37:02Z
- Modified
- 2026-09-10T12:25:27Z
- Severity
-
- 10.0 (Critical) CVSS_V3 - CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:C/C:H/I:H/A:H CVSS Calculator
- Summary
- Xinference: Remote code execution via unsafe `eval()` in Llama3 tool-call parsing
- Details
-
Xinference is an inference API for running open-source, speech, and multimodal models. In 2.5.0 and earlier, Xinference passes attacker-influenced Llama3 tool-call output to eval() in xinference/model/llm/tool_parsers/llama3_tool_parser.py and xinference/model/llm/utils.py. Requests to /v1/chat/completions with a tools field flow through xinference/api/restful_api.py, xinference/model/llm/transformers/core.py, handle_chat_result_non_streaming(), and _post_process_completion() before extract_tool_calls() or _eval_llama3_chat_arguments() evaluates the model-generated Python expression. An unauthenticated remote attacker can influence that output through a crafted prompt and execute commands in the Xinference server process context. This issue is fixed in version 2.7.0.
- Database specific
{ "cna_assigner": "GitHub_M", "cwe_ids": [ "CWE-95" ], "osv_generated_from": "https://github.com/CVEProject/cvelistV5/tree/main/cves/2026/61xxx/CVE-2026-61539.json" }- References
-
- https://github.com/xorbitsai/inference/releases/tag/v2.7.0
- https://github.com/CVEProject/cvelistV5/tree/main/cves/2026/61xxx/CVE-2026-61539.json
- https://github.com/xorbitsai/inference/security/advisories/GHSA-x2rj-828p-hx9m
- https://nvd.nist.gov/vuln/detail/CVE-2026-61539
- https://github.com/xorbitsai/inference/commit/1b3d220f342ce68d34cec4586d9409d457dadc42
- https://github.com/xorbitsai/inference/pull/4786