sierra-interactive

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

• Third-party content exposure detected (high risk: 0.80). The SKILL.md explicitly instructs using Membrane to run actions and to proxy arbitrary requests (e.g., "membrane action run ..." and "membrane request CONNECTION_ID /path/to/endpoint"), which causes the agent to fetch and interpret data from the Sierra Interactive API (leads, messages, website visitor content and other user-generated CRM data) — untrusted third-party content that could influence agent behavior.