a11y-ally

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

• Third-party content exposure detected (high risk: 0.90). The skill explicitly navigates to and scrapes arbitrary TARGET_URL webpages (see "STEP 1: BROWSER AUTOMATION
• Content Fetching" with vibium/page.goto and page.evaluate), downloads and analyzes third‑party videos/frames (STEP 7), and requires the LLM to read and infer HTML/image/video content to generate remediation and drive follow-up actions (STEP 3 and STEP 7), so untrusted public web content can directly influence agent decisions.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill injects and loads axe-core at runtime via the CDN URL https://cdnjs.cloudflare.com/ajax/libs/axe-core/4.10.2/axe.min.js (vibium eval appends a script tag and awaits s.onload), which executes remote JavaScript in the page and is a required dependency for the scan, so it meets the criteria for a runtime external code dependency.