website-to-hyperframes

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

• Third-party content exposure detected (medium risk: 0.30). Step 1’s capture/extraction path ingests outsider-authored free text by crawling a user-supplied URL and reading extracted/visible-text.txt (and related extracted files) to summarize and drive later steps.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill explicitly fetches arbitrary website content at runtime via the capture command (npx hyperframes capture — e.g., "https://...") which is then read and injected into the agent workflow to produce DESIGN.md/SCRIPT.md/STORYBOARD (thus external content directly controls agent outputs), and the techniques include CDN script URLs that would execute remote code at render time (e.g., https://cdn.jsdelivr.net/npm/@dotlottie/player-component@2.7.12/dist/dotlottie-player.js and https://cdn.jsdelivr.net/npm/gsap@3.14.2/dist/MotionPathPlugin.min.js).