software-accessibility

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

• Third-party content exposure detected (medium risk: 0.30). The skill’s workflow describes running axe-core/Lighthouse/Pa11y against a user-supplied URL (e.g., scripts/run_axe.sh takes TARGET_URL and fetches the page to extract/scan rendered text and markup), so an outsider can submit free text content via a web page that the automation ingests without selecting a specific trusted item first.

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

• Potentially malicious external URL detected (high risk: 0.80). scripts/run_axe.sh invokes npx to run @axe-core/cli (fetching and executing remote code at runtime), and the skill references the axe-core repository URL https://github.com/dequelabs/axe-core, indicating a required runtime fetch-and-execute dependency.