audit-website-aeo

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

• Third-party content exposure detected (high risk: 1.00). The skill's crawler (scripts/aeo-audit.mjs) fetches and crawls a live, user-supplied website and the SKILL.md Step 4 explicitly instructs the agent to read the pagesForReview fields (textExcerpt, headings, metadata) from that untrusted third‑party site and use them to score and drive the report and recommendations, which can materially change tool actions—creating a clear avenue for indirect prompt injection.

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

• Potentially malicious external URL detected (high risk: 0.80). The skill fetches and injects arbitrary pages from the user-supplied target site at runtime (including /llms.txt, e.g. https://example.com/llms.txt), and that fetched content is placed into pagesForReview and used directly by the agent to drive its scoring—so remote content can be crafted to influence the model's prompts/evaluation.