dixa

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 proxy requests against the Dixa API (e.g., "membrane request CONNECTION_ID /path/to/endpoint") to retrieve conversation messages and other user-generated content (email/chat/social media) that the agent is expected to read and act on, so untrusted third-party content could influence behavior.