mindbreeze

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

• Third-party content exposure detected (high risk: 0.80). The SKILL.md shows the agent uses Membrane to run searches and proxy requests to Mindbreeze (e.g., membrane action run, membrane request) and explicitly lists entity types like "Web site", "Blog Post", "Forum Post", and "Social Media Post", indicating the agent will fetch and read untrusted/public third‑party content as part of its workflow.