CVE-2026-76841

HIGHPre-NVD 8.88.8—
EchelonGraph scoreHIGH confidence

Score 8.8 from GitHub Security Advisory (severity: HIGH) published 2026-08-24. the CNA's CVSS baseline 8.8; sources differ by 0.0.

Triggered by: GitHub Security Advisory CVSS
Sources: cna:vulncheck, epss, ghsa
Trending — 4 sources updated this week
8.8EG
EchelonGraph verdictPlan a fixSerious severity, but no confirmed exploitation yet.
  • High severity, but no confirmed exploitation yet
CISA-KEV: Not listedEPSS PROB: 1%CVSS: 8.8Exploit: None knownExposed: 0

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

Xinference loads models with Hugging Face remote code execution unconditionally enabled, and before version 2.12.0 exposes no setting to disable it. Six loader call sites pass trust_remote_code=True as a literal or as an unconditional default: RerankModel._get_tokenizer in xinference/model/rerank/core.py, SentenceTransformerRerankModel.load in xinference/model/rerank/sentence_transformers/core.py, SentenceTransformerEmbeddingModel.load in xinference/model/embedding/sentence_transformers/core.py, FlagEmbeddingModel.load in xinference/model/embedding/flag/core.py, and two sites in xinference/model/llm/transformers/core.py where PytorchModel._sanitize_model_config and PytorchModel._get_components default the value to True. Because a caller with model launch access can register a model whose type is unknown and supply an arbitrary model path, the server reaches _auto_detect_type and then AutoTokenizer.from_pretrained, which imports and executes Python declared by the model directory's own tokenizer_config.json auto_map, running attacker-supplied code with the privileges of the worker process. Version 2.12.0 gates every site behind allow_trust_remote_code and the XINFERENCE_TRUST_REMOTE_CODE setting, permitting remote code only for bundled built-in models.

CVSS v3
8.8
EG Score
8.8(high)
EG Risk
40(Track)
EG Risk 40/100SSVC: Track

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
Severity88% × 45%
Exploitation1% × 40%
Automatability0% × 15%
Action: Routine — remediate on your standard cadence.
EPSS PROB
1%
EPSS %ILE
51%
KEV
Not listed

Published

August 24, 2026

Last Modified

August 29, 2026

Advisory Details (5)

Auto-updated Aug 24, 2026
Patch available. Sources: github_pr.
generic

Xinference through 2.11.0 Remote Code Execution via Hardcoded trust_remote_code in Model Loaders | Advisories | VulnCheck

https://www.vulncheck.com/advisories/xinference-through-remote-code-execution-via-hardcoded-trust-remote-code-in-model-loaders
generic

inference/xinference/model/rerank/core.py at v2.11.0 · xorbitsai/inference · GitHub

https://github.com/xorbitsai/inference/blob/v2.11.0/xinference/model/rerank/core.py
github_pr Patch Available

feat: gate model remote-code execution behind a config switch

Patch available: xorbitsai/inference v2.12.0 (PR #5027 merged 2026-06-21)

https://github.com/xorbitsai/inference/pull/5027
generic

Remote Code Execution via Hardcoded `trust_remote_code=True` in Model Loaders · Issue #5023 · xorbitsai/inference · GitHub

https://github.com/xorbitsai/inference/issues/5023
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

Vendor Advisories for CVE-2026-76841(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 28× in last 7d / 28× 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. 2026-08-30 01:22 UTCEPSS rescore
  2. 2026-08-29 23:07 UTCGHSA enrichment
  3. 2026-08-29 12:03 UTCEG score recompute
  4. 2026-08-29 12:03 UTCGHSA enrichment
  5. 2026-08-29 00:58 UTCEG score recompute
  6. 2026-08-29 00:58 UTCGHSA enrichment
  7. 2026-08-28 21:42 UTCEPSS rescore
  8. 2026-08-28 13:51 UTCGHSA enrichment
  9. 2026-08-28 02:43 UTCGHSA enrichment
  10. 2026-08-27 15:40 UTCEG score recompute
  11. 2026-08-27 15:40 UTCGHSA enrichment
  12. 2026-08-27 14:25 UTCEPSS rescore
  13. 2026-08-27 04:30 UTCGHSA enrichment
  14. 2026-08-26 17:24 UTCEG score recompute
  15. 2026-08-26 17:24 UTCGHSA enrichment
  16. 2026-08-26 16:23 UTCEG score recompute
  17. 2026-08-26 16:23 UTCGHSA enrichment
  18. 2026-08-26 14:47 UTCEPSS rescore
  19. 2026-08-26 10:38 UTCGHSA enrichment
  20. 2026-08-25 23:34 UTCEG score recompute
  21. 2026-08-25 23:34 UTCGHSA enrichment
  22. 2026-08-25 13:49 UTCEPSS rescore
  23. 2026-08-25 12:31 UTCGHSA enrichment
  24. 2026-08-25 01:27 UTCEG score recompute
  25. 2026-08-25 01:27 UTCGHSA enrichment
Show 3 more
  1. 2026-08-24 14:24 UTCEG score recompute
  2. 2026-08-24 13:22 UTCEG score recompute
  3. 2026-08-24 13:22 UTCMITRE cvelistV5first tracked

Related CVEs(same CWE)

Frequently asked(5)

What is CVE-2026-76841?
CVE-2026-76841 is a high vulnerability published on August 24, 2026. Xinference loads models with Hugging Face remote code execution unconditionally enabled, and before version 2.12.0 exposes no setting to disable it. Six loader call sites pass trustremotecode=True as a literal or as an unconditional default: RerankModel.gettokenizer in…
When was CVE-2026-76841 disclosed?
CVE-2026-76841 was first published in the National Vulnerability Database on August 24, 2026, with the most recent update on August 29, 2026. EchelonGraph re-ingests CVE updates from NVD on a 2-hour cycle, so this page reflects the latest published state.
Is CVE-2026-76841 actively exploited?
CVE-2026-76841 is not currently on CISA's Known Exploited Vulnerabilities catalog. FIRST EPSS estimates a 1% probability of exploitation in the next 30 days, which ranks it in the top 49.4% of all scored CVEs.
What is the CVSS score of CVE-2026-76841?
CVE-2026-76841 has a CVSS v4.0 base score of 8.8 (CNA self-assessment; NVD's own analysis pending).
How do I remediate CVE-2026-76841?
Patch to the fixed version published by the affected vendor. Where vendor advisories exist for CVE-2026-76841, 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-2026-76841

Explore →

Is Your Infrastructure Affected by CVE-2026-76841?

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

Start Free Scan →Browse All CVEs