pixiebrix

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

• Third-party content exposure detected (high risk: 0.80). This skill's SKILL.md explicitly instructs the agent to fetch and run actions against a third-party PixieBrix API via Membrane (see the "Running actions" and "Proxy requests" sections showing membrane action run and membrane request), which can return user-generated/untrusted content (e.g., comments, shared links, files) that the agent is expected to read and could influence subsequent actions, enabling indirect prompt injection.