picky-assist

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

• Third-party content exposure detected (high risk: 0.80). The skill's workflow instructs the agent to fetch and interact with external Picky Assist data via Membrane commands (e.g., "membrane action run" and "membrane request CONNECTION_ID /path/to/endpoint"), which will return user-generated/untrusted content like conversations and contacts that the agent is expected to read and act on.