CVE-2026-100647
Published:September 26, 2026
Updated:October 10, 2026
vLLM versions before 0.29.0 contain a denial-of-service vulnerability in the cache_salt parameter accepted on OpenAI-compatible and Anthropic API endpoints, which lacks maximum length validation and is processed on the single EngineCore scheduler thread. Unauthenticated attackers can send HTTP requests with multi-hundred-megabyte salt values that trigger expensive pickle serialization and SHA-256 hashing, stalling the scheduler thread and denying service to all concurrent requests.
Affected Packages
vllm (CONDA):
Affected version(s) >=0.9.2 <0.29.0Fix Suggestion:
Update to version 0.29.0https://github.com/vllm-project/vllm.git (GITHUB):
Affected version(s) >=v0.9.0 <v0.29.0Fix Suggestion:
Update to version v0.29.0vllm (PYTHON):
Affected version(s) >=0.9.0 <0.29.0Fix Suggestion:
Update to version 0.29.0Related Resources (4)
Do you need more information?
Contact UsCVSS v4
Base Score:
6.9
Attack Vector
NETWORK
Attack Complexity
LOW
Attack Requirements
NONE
Privileges Required
NONE
User Interaction
NONE
Vulnerable System Confidentiality
NONE
Vulnerable System Integrity
NONE
Vulnerable System Availability
LOW
Subsequent System Confidentiality
NONE
Subsequent System Integrity
NONE
Subsequent System Availability
NONE
CVSS v3
Base Score:
5.3
Attack Vector
NETWORK
Attack Complexity
LOW
Privileges Required
NONE
User Interaction
NONE
Scope
UNCHANGED
Confidentiality
NONE
Integrity
NONE
Availability
LOW
Weakness Type (CWE)
Improper Input Validation
EPSS
Base Score:
0.35