modelopt-benchmark-model-kernels

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (low risk: 0.10). The required workflow reads a user-chosen model (local directory, config.json, or Hub ID) and builds it on meta tensors, so outsider-authored strings from that model's config or module names can reach the agent only through a model the user selected, which is a low-exposure path.