one-ai

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

• Third-party content exposure detected (high risk: 0.80). The skill's main workflow uses Membrane to call the One AI connector and proxy arbitrary One AI endpoints (e.g., "membrane action run" and "membrane request CONNECTION_ID /path/to/endpoint") to access Documents/Workspaces, so the agent will fetch and interpret external/user-generated content from the One AI service that could contain instructions.