EUVD Vulnerability Catalog

EUVD-2026-64146

Severity: CRITICAL Base Score: 10 CVSS Version: 3.1

Vulnerability Description

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.

CVSS Vector Analysis

Below is the complete, human-readable breakdown of the CVSS metric string: CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:C/C:H/I:H/A:H

▪ Attack Vector (AV): Network

▪ Attack Complexity (AC): Low

▪ Privileges Required (PR): None

▪ User Interaction (UI): None

▪ Scope (S): Changed

▪ Confidentiality Impact (C): High

▪ Integrity Impact (I): High

▪ Availability Impact (A): High

Affected Vendors & Systems

Vendor xorbitsai

References & Advisory Links

Metadata Profile

Database Authority

European Union Agency for Cybersecurity (ENISA) EUVD

Assigner

GitHub_M

EPSS Probability

Known Aliases

CVE-2026-61539

Published On

2026-08-21

Last Updated

2026-08-21