covalent

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

• Third-party content exposure detected (high risk: 0.80). The skill explicitly exposes and returns external, user-supplied content (e.g., the "Get NFT Metadata" action that "Returns external metadata for an NFT token") via the Covalent/Membrane integration, meaning the agent will fetch and read untrusted public third‑party metadata which could influence subsequent decisions or actions.