site-launch-checklist

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

• Third-party content exposure detected (high risk: 1.00). The skill's required workflow (see "Verification tools", Phase 5/6, and the weekly SEO agent) explicitly instructs the agent to fetch and parse public third-party resources and services (e.g., robots.txt, sitemap.xml, page HTML/JSON-LD, Google Trends, Ahrefs, securityheaders.com, opengraph.xyz, productrank.ai) via curl/WebFetch/Chrome extension and uses those results to decide policies and next actions (AI-scraper rules, installs, fixes), which exposes it to untrusted third-party content that could carry indirect prompt-injection.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill explicitly instructs running runtime installs like "npx skills add owner/repo" and even shows a direct GitHub install (e.g., https://github.com/blader/humanizer), which will fetch and install remote skill code at runtime that can execute and control agent behavior, so this is a runtime external dependency that can run remote code.