revealjs

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

• Third-party content exposure detected (medium risk: 0.30). The skill “revealjs” generates reveal.js single-file HTML where slide content, including speaker notes and (optionally) external markdown references, is derived from the user’s request; the runtime path ingests that user-authored free text directly into the generated HTML.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.90). The skill generates HTML that, at runtime, loads and executes required reveal.js JavaScript/CSS from the jsDelivr CDN (e.g. https://cdn.jsdelivr.net/npm/reveal.js@5/dist/reveal.js), meaning remote code is fetched and executed and the skill relies on those CDN assets for functionality.