website-rebuild

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

• Third-party content exposure detected (high risk: 0.85). The skill’s runtime workflow ingests outsider-authored free text by having the user-supplied URL’s pages be opened and scraped (Phase 1 “Open the URL in Chrome… Visit every page” and Phase 2 “Write and run a scraping script… Extract all text… from every page”), which is then placed into LLM context for brand/content/SEO/copy planning.

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

• Potentially malicious external URL detected (high risk: 0.80). During QA the skill's preview_eval snippet dynamically loads and executes remote JavaScript from the CDN URL https://cdnjs.cloudflare.com/ajax/libs/axe-core/4.9.1/axe.min.js, which is a runtime fetch that executes external code in the agent/browser environment.