google-cloud-secret-manager-js

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

• Third-party content exposure detected (high risk: 0.85). The skill’s runtime path is calling Google Secret Manager via the official client (client.accessSecretVersion({ name })), which returns secret payload bytes that are decoded into UTF-8 and inserted into the agent/app context (i.e., outsider-authored secret values are ingested as free text).