Evidence · DiggingBeagle record
vLLM Responses API template trust-policy bypass, CVE-2026-115918
Advisory showing a distinct OpenAI-compatible endpoint missed request-template validation.
- Source role
- primary disclosure
Evidence record
Advisory showing a distinct OpenAI-compatible endpoint missed request-template validation.
Claim-level citations (2)
- supportsvLLM Responses endpoint accepted chat-template kwargs despite a disabled trust flag: The vLLM Responses API advisory reports that nested chat_template_kwargs escaped the same trust_request_chat_template restriction enforced by other chat routes, allowing forbidden request templates to reach rendering.
GHSA-x567-49g5-p4cf: trust-request-chat-template, /v1/responses and affected versions
- supportsvLLM Responses endpoint accepted chat-template kwargs despite a disabled trust flag: vLLM applied trust-request-chat-template validation to Chat endpoints but failed to apply the same rejection to nested template kwargs arriving through its Responses API. The affected service must expose the OpenAI-compatible Responses route and pass caller-controlled chat_template_kwargs while request templates are disabled in configuration. A client could get a forbidden template evaluated and potentially combine the bypass with the separate Jinja CPU exhaustion mechanism.
GHSA-x567-49g5-p4cf: trust-request-chat-template, /v1/responses and affected versions