CVE-2026-115917

vLLM is an inference and serving engine for large language models. Prior to 0.32.0, the OpenAI-compatible server configures Starlette CORSMiddleware with allowed_origins set to a wildcard by default. A malicious web page can use a victim's browser to reach an unauthenticated vLLM server on localhost, an internal network, or a VPN, call endpoints including GET /v1/models and POST /v1/chat/completions, consume GPU resources, and read responses. Cookie-based cross-site request forgery does not apply because credentials are disabled by default, and deployments requiring bearer authentication are exposed only when the caller already has a usable token. This issue is fixed in version 0.32.0.

• https://github.com/vllm-project/vllm/commit/fcae5c98870ff2c20659731dce7ba3a3ff0d48aa
• https://github.com/vllm-project/vllm/pull/61009
• https://github.com/vllm-project/vllm/releases/tag/v0.32.0
• https://github.com/vllm-project/vllm/security/advisories/GHSA-vvpg-rrww-wpmh

• https://github.com/vllm-project/vllm/commit/fcae5c98870ff2c20659731dce7ba3a3ff0d48aa
• https://github.com/vllm-project/vllm/pull/61009
• https://github.com/vllm-project/vllm/releases/tag/v0.32.0
• https://github.com/vllm-project/vllm/security/advisories/GHSA-vvpg-rrww-wpmh

• UNSCOREDCVE-2026-107858PoC
• UNSCOREDCVE-2026-108372PoC
• MEDIUMCVE-2026-111922PoC
• HIGHCVE-2026-116748PoC
• LOWCVE-2026-3303
• MEDIUMCVE-2026-55842PoC
• LOWCVE-2026-56595
• UNSCOREDCVE-2026-66070PoC

Related by shared AI tags and CWE weakness class. Browse the full archive.