Tribalknowledge

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

• Third-party content exposure detected (high risk: 0.75). The skill’s runtime workflow ingests user-provided “knowledge capture” text (e.g., from “Capture what Dave knows…” / interview answers) into the LLM context; since that free-form content is authored by machinists other than the operating user, it is outsider-authored free text via the capture/interview input path.