CriticalVulnerabilityLLM-specific

CVE-2026-63764: lmdeploy's OpenAI-compatible API server contains a server-side request forgery vulnerability that allows…

Source: NVD(opens in a new tab)
Identifier: CVE-2026-63764
Published: July 21, 2026
Record updated: October 9, 2026

Known exploitation: Not listed in the CISA Known Exploited Vulnerabilities catalog at the last check.
EPSS: 0.5%

Summary

lmdeploy's OpenAI-compatible API server contains a server-side request forgery vulnerability (CVE-2026-63764) that unauthenticated attackers can reach through a POST to the chat completions endpoint. A crafted image_url pointing to an attacker-controlled server returns an HTTP 302 redirect to internal addresses such as loopback or instance-metadata endpoints. The initial URL safety check passes because redirects are followed without re-validating each hop through the safety guard.

Mitigation

The source does not state a fix yet. Check the original advisory for updates.

Details

Scores and exploitation

CVSS score: 9.3(critical)
CVSS vector: CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:C/C:H/I:L/A:N
Attack vector: network
Attack complexity: low
Privileges required: none
User interaction: none
EPSS: 0.5%Estimated probability that this CVE is exploited in the next 30 days, from the FIRST Exploit Prediction Scoring System. Read on November 12, 2026.
Disclosure date: July 21, 2026

Weaknesses and techniques

Weakness (CWE): CWE-918
Attack pattern (CAPEC): CAPEC-664

Classification

Attack type: Other
Attack sophistication: Trivial
AI component targeted: API
Impact: Confidentiality
Labels: security

How this record was produced

Summaries, mitigations and labels on this page are generated automatically from the source. The linked source is authoritative.

Original source: https://nvd.nist.gov/vuln/detail/CVE-2026-63764

First tracked: July 22, 2026 at 02:07 AM ET, from NVD/CVE Database

Classifier: claude-haiku-5-5, prompt v4

Classifier confidence: 93% (self-assessed)

Summary

Mitigation

Related items