double

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

• Third-party content exposure detected (high risk: 0.70). The SKILL.md explicitly exposes actions that fetch and read user-generated third-party content (e.g., "List Posts", "Get Post", "List Comments") and allows arbitrary proxied API requests via "membrane request", so the agent will ingest external post/comment content that could influence subsequent actions.