LMDeploy is a toolkit for compressing, deploying, and serving large language models. Versions prior to 0.12.3 have a Server-Side Request Forgery (SSRF) vulnerability in LMDeploy's vision-language module. The `load_image()` function in `lmdeploy/vl/utils.py` fetches arbitrary URLs without validating internal/private IP addresses, allowing attackers to access cloud metadata services, internal networks, and sensitive resources. Version 0.12.3 patches the issue.
References
| Link | Resource |
|---|---|
| https://github.com/InternLM/lmdeploy/commit/71d64a339edb901e9005358e0633fbbab367d626 | Patch |
| https://github.com/InternLM/lmdeploy/pull/4447 | Issue Tracking Patch |
| https://github.com/InternLM/lmdeploy/releases/tag/v0.12.3 | Release Notes |
| https://github.com/InternLM/lmdeploy/security/advisories/GHSA-6w67-hwm5-92mq | Exploit Mitigation Vendor Advisory |
| https://github.com/InternLM/lmdeploy/security/advisories/GHSA-6w67-hwm5-92mq | Exploit Mitigation Vendor Advisory |
Configurations
History
No history.
Information
Published : 2026-04-20 21:16
Updated : 2026-06-17 10:37
NVD link : CVE-2026-33626
Mitre link : CVE-2026-33626
CVE.ORG link : CVE-2026-33626
JSON object : View
Products Affected
internlm
- lmdeploy
CWE
CWE-918
Server-Side Request Forgery (SSRF)
