sifter

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

• Third-party content exposure detected (high risk: 0.80). The skill's SKILL.md explicitly instructs using the Membrane CLI to connect to Sifter and to list/run actions that read and return Sifter data (e.g., Issues and Comments under "Sifter Overview" and the "action list"/"action run" commands), which are third-party, user-generated contents the agent will fetch and interpret at runtime.