CVE-2026-85668

HIGHPre-NVD 7.57.5—
EchelonGraph scoreHIGH confidence

Score 7.5 from GitHub Security Advisory (severity: HIGH) published 2026-09-04. a secondary CVSS source baseline 7.5; sources differ by 0.0.

Triggered by: GitHub Security Advisory CVSS
Sources: epss, ghsa, secondary
Trending — 3 sources updated this weekElevated
7.5EG
EchelonGraph verdictPlan a fixSerious severity, but no confirmed exploitation yet.
  • High severity, but no confirmed exploitation yet
CISA-KEV: Not listedEPSS PROB: 0%CVSS: 7.5Exploit: Elevated riskExposed: 0

No vendor fix yet — apply a workaround or compensating control (WAF / firewall / segmentation) and watch for a patch.

Xinference (affected commit 4a94832, v3.x) contains an unauthenticated arbitrary-path file read vulnerability in the POST /v1/models/llm/auto-register endpoint, which accepts a caller-supplied model_path parameter without authentication or path confinement. The endpoint reads and parses config.json, tokenizer_config.json, and chat_template.jinja files at the supplied path and reflects the parsed content back to the caller, allowing an unauthenticated attacker to probe the server filesystem and extract content of files with those names in any directory.

CVSS v3
7.5
EG Score
7.5(high)
EG Risk
65(Attend)
EG Risk 65/100SSVC: Attend

EG Risk is EchelonGraph's 0–100 priority score: it fuses intrinsic severity with real-world exploitation and automatability so you can rank equal-severity CVEs and fix the most dangerous first. Higher = act sooner. Distinct from the 0–10 EG Score (severity).

How it’s computed
Severity75% × 45%
Exploitation40% × 40%
Automatability100% × 15%
Action: Remediate soon — notable exploitation risk.
EPSS PROB
0%
EPSS %ILE
38%
KEV
Not listed

Published

September 4, 2026

Last Modified

September 4, 2026

Advisory Details (4)

Auto-updated Sep 4, 2026
No patch confirmed yet.
generic

Xinference 3.3.0 Unauthenticated Arbitrary-Path File Read via /v1/models/llm/auto-register | Advisories | VulnCheck

https://www.vulncheck.com/advisories/xinference-3.3.0-unauthenticated-arbitrary-path-file-read-via-v1-models-llm-auto-register
generic

inference/xinference/model/llm/config_parser.py at v3.3.0 · xorbitsai/inference · GitHub

https://github.com/xorbitsai/inference/blob/v3.3.0/xinference/model/llm/config_parser.py
generic

GitHub - xorbitsai/inference: Swap GPT for any LLM by changing a single line of code. Xinference lets you run open-source, speech, and multimodal models on cloud, on-prem, or your laptop — all through one unified, production-ready inference API. · GitHub

https://github.com/xorbitsai/inference
generic

Unauthenticated arbitrary-path file read (partial) via /v1/models/llm/auto-register, plus unauthenticated arbitrary module import via flexible launcher · Issue #5176 · xorbitsai/inference · GitHub

https://github.com/xorbitsai/inference/issues/5176

Vendor Advisories for CVE-2026-85668(1)

These vendors published their own advisory mentioning this CVE — often with vendor-specific remediation steps + affected product lists not in NVD.

Weakness Classification(1)

MITRE Common Weakness Enumeration — the root-cause categories this CVE belongs to.

Data Freshness Timeline

(refreshed 26× in last 7d / 58× in last 30d)

Each row is a source pipeline that fetched or updated this CVE on that date, with what changed. For example, "NVD update" means NVD published or revised its analysis for this CVE; "MITRE cvelistV5" means we ingested or refreshed it from the CNA feed. Most recent first.

  1. 2026-09-19 19:27 UTCEG score recompute
  2. 2026-09-19 19:27 UTCGHSA enrichment
  3. 2026-09-19 06:34 UTCEG score recompute
  4. 2026-09-19 06:34 UTCGHSA enrichment
  5. 2026-09-18 19:28 UTCEPSS rescore
  6. 2026-09-18 17:39 UTCGHSA enrichment
  7. 2026-09-18 04:46 UTCEG score recompute
  8. 2026-09-18 04:46 UTCGHSA enrichment
  9. 2026-09-17 19:32 UTCEPSS rescore
  10. 2026-09-17 15:53 UTCGHSA enrichment
  11. 2026-09-17 03:00 UTCEG score recompute
  12. 2026-09-17 03:00 UTCGHSA enrichment
  13. 2026-09-16 14:09 UTCEPSS rescore
  14. 2026-09-16 13:57 UTCEG score recompute
  15. 2026-09-16 13:57 UTCGHSA enrichment
  16. 2026-09-16 05:15 UTCEPSS rescore
  17. 2026-09-16 00:42 UTCGHSA enrichment
  18. 2026-09-15 11:42 UTCEG score recompute
  19. 2026-09-15 11:42 UTCGHSA enrichment
  20. 2026-09-15 03:12 UTCEPSS rescore
  21. 2026-09-15 03:12 UTCEPSS rescore
  22. 2026-09-14 22:49 UTCGHSA enrichment
  23. 2026-09-14 09:55 UTCGHSA enrichment
  24. 2026-09-13 20:58 UTCEG score recompute
  25. 2026-09-13 20:58 UTCGHSA enrichment
Show 33 more
  1. 2026-09-13 08:04 UTCGHSA enrichment
  2. 2026-09-12 19:12 UTCEG score recompute
  3. 2026-09-12 19:12 UTCGHSA enrichment
  4. 2026-09-12 06:18 UTCGHSA enrichment
  5. 2026-09-11 17:26 UTCEG score recompute
  6. 2026-09-11 17:26 UTCGHSA enrichment
  7. 2026-09-11 14:53 UTCEPSS rescore
  8. 2026-09-11 14:53 UTCEPSS rescore
  9. 2026-09-11 03:59 UTCGHSA enrichment
  10. 2026-09-10 15:06 UTCEG score recompute
  11. 2026-09-10 15:06 UTCGHSA enrichment
  12. 2026-09-10 02:13 UTCGHSA enrichment
  13. 2026-09-09 13:20 UTCGHSA enrichment
  14. 2026-09-09 00:22 UTCEG score recompute
  15. 2026-09-09 00:22 UTCGHSA enrichment
  16. 2026-09-08 22:01 UTCEPSS rescore
  17. 2026-09-08 10:59 UTCGHSA enrichment
  18. 2026-09-07 21:54 UTCEG score recompute
  19. 2026-09-07 21:54 UTCGHSA enrichment
  20. 2026-09-07 09:01 UTCGHSA enrichment
  21. 2026-09-06 20:07 UTCEG score recompute
  22. 2026-09-06 20:06 UTCGHSA enrichment
  23. 2026-09-06 13:48 UTCEPSS rescore
  24. 2026-09-06 07:09 UTCGHSA enrichment
  25. 2026-09-05 18:16 UTCEG score recompute
  26. 2026-09-05 18:16 UTCGHSA enrichment
  27. 2026-09-05 15:31 UTCEPSS rescore
  28. 2026-09-05 05:24 UTCGHSA enrichment
  29. 2026-09-04 16:30 UTCEG score recompute
  30. 2026-09-04 16:30 UTCGHSA enrichment
  31. 2026-09-04 15:23 UTCEG score recompute
  32. 2026-09-04 14:45 UTCEG score recompute
  33. 2026-09-04 14:44 UTCMITRE cvelistV5first tracked

Frequently asked(5)

What is CVE-2026-85668?
CVE-2026-85668 is a high vulnerability published on September 4, 2026. Xinference (affected commit 4a94832, v3.x) contains an unauthenticated arbitrary-path file read vulnerability in the POST /v1/models/llm/auto-register endpoint, which accepts a caller-supplied modelpath parameter without authentication or path confinement. The endpoint reads and parses config.json,…
When was CVE-2026-85668 disclosed?
CVE-2026-85668 was first published in the National Vulnerability Database on September 4, 2026. EchelonGraph re-ingests CVE updates from NVD on a 2-hour cycle, so this page reflects the latest published state.
Is CVE-2026-85668 actively exploited?
CVE-2026-85668 is not currently on CISA's Known Exploited Vulnerabilities catalog. FIRST EPSS estimates a 0% probability of exploitation in the next 30 days, which ranks it in the top 61.6% of all scored CVEs.
What is the CVSS score of CVE-2026-85668?
CVE-2026-85668 has a CVSS v3 base score of 7.5 (NVD).
How do I remediate CVE-2026-85668?
Patch to the fixed version published by the affected vendor. Where vendor advisories exist for CVE-2026-85668, EchelonGraph cross-links them in the Vendor Advisories panel below — those typically contain the canonical remediation steps, fixed version numbers, and any vendor-specific mitigations.

Dependency Blast Radius

Explore the affected products and dependency analysis for CVE-2026-85668

Explore →

Is Your Infrastructure Affected by CVE-2026-85668?

EchelonGraph automatically scans your cloud infrastructure and maps CVE exposure using blast radius analysis.

Start Free Scan →Browse All CVEs