non-commodity-content

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

• Third-party content exposure detected (high risk: 1.00). The skill's Step 1 explicitly instructs "If the input is a URL, WebFetch it" and requires the agent to fetch and read arbitrary public URLs as part of its core workflow, which exposes it to untrusted third‑party web content that could contain injected instructions.

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

• Potentially malicious external URL detected (high risk: 0.80). The skill explicitly says "If the input is a URL, WebFetch it" (e.g., https://mybrand.com/blog/seo-tips), meaning user-supplied page content is fetched at runtime and injected into the agent's prompt/analysis flow, so external content can directly control outputs.