validate-block-markup

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

• Third-party content exposure detected (medium risk: 0.30). The skill runtime scripts/validate-block-markup.cjs ingests outsider-authored free text directly from user-supplied paths or stdin (via readInputs), then parses/validates it in validateBlockMarkup without any prior selection of a specific trusted item.

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

• Potentially malicious external URL detected (high risk: 1.00). The skill bootstraps at first run by running "npm ci" (SKILL.md:24-28) which fetches and installs the pinned Node packages from the npm registry (e.g. https://registry.npmjs.org/jsdom/-/jsdom-26.1.0.tgz), bringing remote JavaScript code into the skill that will be loaded and executed at runtime.