grillkit

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

• Third-party content exposure detected (medium risk: 0.30). Source is the user’s own free-text idea passed into the grillkit skill via the skill’s “grill” prompt/workflow (“before the first question, restate the subject” and then repeatedly asks decisions), so the runtime necessarily ingests outsider-authored content that the user supplies.