fiction-writer

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

• Third-party content exposure detected (high risk: 1.00). The skill explicitly instructs the agent to use the Browser to search the open web, Google Images, and YouTube and to include real external URLs as references and images (see "Recherche et inspiration", "Important pour les références", and the image-search instructions), so it will fetch and read untrusted public third‑party content that could carry indirect prompt injection.