kanban

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

• Third-party content exposure detected (medium risk: 0.65). The required runtime workflow for the connector skill (kb_health/kb_summary/etc.) POSTs to a board URL (community or localhost) and ingests the board’s JSON fields (e.g., task titles/descriptions/participants) into the agent’s tool outputs, which are then available as LLM context; if the board URL targets another user’s published board, that content is outsider-authored free text.