Internal-service data disclosure via unchecked HTTP redirects
Published Jul 21, 2026 · Updated Jul 21, 2026
SSRF in InternLM LMDeploy through 0.14.0 allows remote attackers to read internal service content through a redirected image_url. The _load_http_url function in lmdeploy/vl/media/connection.py validates only the initial host, then follows HTTP redirects without applying the private-IP guard to each destination. The OpenAI-compatible vision server binds to all interfaces without authentication by default, so a crafted chat-completions request can disclose internal-service or cloud-metadata content through the model pipeline.
Summary
What happened
SSRF in InternLM LMDeploy through 0.14.0 allows remote attackers to read internal service content through a redirected image_url. The _load_http_url function in lmdeploy/vl/media/connection.py validates only the initial host, then follows HTTP redirects without applying the private-IP guard to each destination. The OpenAI-compatible vision server binds to all interfaces without authentication by default, so a crafted chat-completions request can disclose internal-service or cloud-metadata content through the model pipeline.
The record
- CVE
- CVE-2026-63764
- Published
- Jul 21, 2026
- Updated
- Jul 21, 2026
- Vendor
- InternLM
- Product
- LMDeploy
- Classifications
- CWE-918, T1552.005, T1190
- Attack vector
- network
- Privileges
- unauthenticated
Timeline
How it unfolded
- Jul 21, 2026CVE publishedPublication date reported by the CVE source.
- Jul 21, 2026Record updatedLatest update available in the CVE record.
Exploitability
Present is not the same as exploitable
Compare your product and version with the public record. A matching version still requires validation against your environment.
Is a vulnerable build present?
Affected versions are unavailable in this record. Check the vendor advisory for version and patch details.
What conditions does exploitation require?
What is affected?
Attacks
What attackers are doing with it
Daily unique IPs observed by Shadowserver honeypots for known exploited vulnerabilities (KEVs). Missing observations do not establish an absence of attacks.
Weakness, pattern, technique
Public exploit references
- Unauthenticated SSRF via image_url reproductionproof of concept · demonstrated
Labels summarize the accepted research assessment. They do not indicate a test against your environment.
Technologies
Your stack
See the directory against your own environment.
Your stack
Check the software in your environment
Book a demo to see how Hinoki identifies affected software and validates exploitability in your environment.
Book a demo