ads-multi

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

• Third-party content exposure detected (high risk: 0.90). The skill's Step 1 "Resolve landing page per language" explicitly says it will "fetch and cache" landing pages and allows users to "provide URLs per language manually", which means the agent will retrieve and analyze arbitrary public landing-page URLs (third‑party web content) and use that content to shape ad copy—exposing it to untrusted external input.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill explicitly fetches and caches user-provided landing page URLs at runtime (the "Landing page" input / resolved variants from .claude/skills/landing-pages/SKILL.md), and that fetched page content is used to inform/generate ad copy (i.e., injected into the model context), so external URLs provided at runtime can directly control prompts.