LMDeploy Vulnerable to Server-Side Request Forgery (SSRF) via Vision-Language Image Loading
90Vexday Risk Score
Patch now. It exploitation observed by VulnCheck and has a working public exploit.
ssvc Actcvss 7.5epss 45%
from disclosure to weapon23 days
Published on NVDApr 20
1st PoC+23d
VulnCheck+2d
exploitation probability
45%top 1% of all CVEs
observed exploitation
yesVulnCheck
1 public exploit(s)
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.
CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:N/A:N
Affected products
InternLM · lmdeploypublic PoCs found — 1
githubgithub.com/rootdirective-sec/CVE-2026-33626-Lab★ 0⚠ Public resources, to assess the exposure of systems you control or are authorized to test. Test only with authorization.