Skip to main content
No exploitation reportedFix available

CVE-2026-61539

Xinference: Remote code execution via unsafe `eval()` in Llama3 tool-call parsing

No exploitation reported by the sources we read.

  • Published2026-08-21
  • Fixed in2.7.0
  • WeaknessCWE-95

Am I affected?

With a domain, we read the version its public headers and pages expose. A distribution build can carry the fix under an older number.

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.

Timeline

dates in UTC
  1. 2026-08-2121:17CVE publishedNVD

Versions per branch

OSV
xinference
BranchAffectedFixed in
2.7< 2.7.02.7.0

as publishedRanges as NVD and OSV publish them; a distribution’s build can be fixed under an older number.

Products declared by the CNA1 product

As GitHub_M, the CNA that assigned this CVE, declared them in the CVE record; versions as written.

  • xorbitsai inference
    Affected
    • = < 2.7.0no fix stated

Source: the CNA record (CVE Program), © The MITRE Corporation,CVE Terms of UseCVE recordupdated 2026-08-21

Fix

  1. Update xinference to 2.7.0

Detection

Nuclei template
none so far
ProjectDiscovery
Exploit-DB
no entry
Exploit-DB
Proof of concept
Metasploit modules
none listed
Metasploit Framework

Sources

  • NVDCVE recordread 2026-10-23
  • CISA-ADPSSVC decisionnone2026-08-24
  • CNACVSS 3.110known since 2026-10-23
  • CNAweakness (CWE)CWE-95known since 2026-10-23
  • FIRST EPSSEPSS score1.2 %2026-11-18
  • OSVaffected packages—
  • GitHub AdvisoriesadvisoryCRITICAL2026-08-21

Identifiers

Technical details

CVSS v3 Vector
CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:C/C:H/I:H/A:H
Published
2026-08-21
Last Modified
2026-10-12
NVD status
Deferred
GitHub Advisory
GHSA-x2rj-828p-hx9m

Frequently asked questions

What is CVE-2026-61539?

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.

Is CVE-2026-61539 actively exploited?

Active exploitation of CVE-2026-61539 has not been confirmed. Its EPSS score was 1.2% on 2026-11-18, the estimated probability of exploitation in the next 30 days.

What is the CVSS score for CVE-2026-61539?

CVE-2026-61539 has a CVSS v3 base score of 10 (CRITICAL severity), with vector string CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:C/C:H/I:H/A:H.

Running xinference? Hear about its next vulnerability

Surveillance reads the versions in your stack — a domain, a lockfile or a list — checks each CVE against the ranges its sources publish, and alerts you when one you run is exploited or enters CISA KEV. A version it cannot read stays « version not verified ».

No credit card required · No monthly cap on website lookups · 250 API requests/month

Sources for this record