ModelScope 1.11.0 to before 1.40.1 contains an unsafe reflection vulnerability in the image-to-3D SyncDreamer pipeline
Source GCVE Database
Not known to be exploited · CVSS 8.5 High : upgrade to 1.40.1 soon.
Data as of 17 Nov 2026, 23:33 UTC
Thin evidence
One scorer or none| CNA (VulnCheck) | CVSS 4.0 | 8.5 |
|---|---|---|
| CNA (VulnCheck) | CVSS 3.1 | 7.8 |
- Patch
- Available – fixed in 1.40.1
CVSS:4.0/AV:L/AC:L/AT:N/PR:N/UI:P/VC:H/VI:H/VA:H/SC:N/SI:N/SA:NExploitation evidence
Disclosed: evidence found.
CVE published 17 Nov 2026
Detectable: no evidence on record.
No public nuclei template indexed
Weaponized: no evidence on record.
No public exploit code indexed in Metasploit or Exploit-DB
Exploited in the wildExploited: no evidence on record.
Not in CISA KEV and no source reports exploitation
Observed by OffSeq MirageSeen by Mirage: no evidence on record.
No attack traffic recorded by OffSeq Mirage sensors
A missing step is not evidence of safety. Sources: cve.org (opens in a new tab), the Metasploit, Exploit-DB and nuclei-templates index as of 17 Nov 2026.
What to do
Upgrade ModelScope to 1.40.1.
Affected software
| Product | Affected | Fixed in |
|---|---|---|
| ModelScopeModelScope | Affected
| Fixed in
|
PyPIModelScopepkg:pypi/ModelScopeOther threats | Affected
| Fixed in
|
Check your own hosts for these packages with threat-finder (opens in a new tab), our free open-source scanner.
AI-suggested
If you cannot patch yet, users should avoid loading untrusted or unauthenticated model configurations, especially those containing syncdreamer.yaml target fields. Implementing allowlists or trust_remote_code checks before loading models can mitigate exploitation risk.
Description
ModelScope versions from 1.11.0 up to but not including 1.40.1 contain an unsafe reflection vulnerability in the image-to-3D SyncDreamer pipeline. This flaw allows remote attackers to execute arbitrary code by crafting a model configuration file that abuses the syncdreamer.yaml target field. The vulnerability arises because instantiate_from_config() imports and instantiates any class with attacker-controlled parameters without an allowlist or trust_remote_code verification, leading to code execution upon loading the malicious model.
References
- nvd.nist.gov/vuln/detail/CVE-2026-130839 (opens in a new tab)
- github.com/modelscope/modelscope/pull/1788 (opens in a new tab)
- github.com/modelscope/modelscope/commit/563133cfbc7af7891498d9928a1b018fa42ce30b (opens in a new tab)
- github.com/modelscope/modelscope/blob/v1.40.0/modelscope/models/cv/image_to_3d/ldm/util.py (opens in a new tab)
- github.com/modelscope/modelscope/blob/v1.40.0/modelscope/pipelines/cv/image_to_3d_pipeline.py (opens in a new tab)
- github.com/modelscope/modelscope/releases/tag/v1.40.1 (opens in a new tab)
- vulncheck.com/advisories/modelscope-1.11.0-to-before-1.40.1-unsafe-reflection-rce-via-syncdreamer-yaml (opens in a new tab)