sender

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

• Third-party content exposure detected (high risk: 0.80). The SKILL.md explicitly tells the agent to fetch and interpret data from the external Sender API via Membrane (e.g., "membrane action list ... --connectionId=CONNECTION_ID" which returns action objects/inputSchema and "membrane request CONNECTION_ID /path/to/endpoint"), so the agent will ingest untrusted, user-generated third‑party content that can influence which actions/requests it runs.