slidev

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

• Third-party content exposure detected (high risk: 1.00). The skill explicitly supports embedding and loading arbitrary open-web content—e.g., iframe layouts with a frontmatter url, the and components, automatic caching of remote images (references/build-remote-assets.md), and PlantUML via a public plantuml server—so the agent will fetch and render untrusted, user-generated third-party content as part of its workflow, enabling indirect prompt injection.