authentication-security

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

• Third-party content exposure detected (medium risk: 0.30). The skill’s runtime workflow ingests outsider-authored free text only via HTTP request inputs (e.g., JWT/access tokens, OAuth2 authorization code/state/error query params) that are presented directly by the caller, and there is no prior “select a specific item” gating step—so an outsider can submit arbitrary token/code/state strings to the endpoints the templates implement.