google-ads-audit

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

• Third-party content exposure detected (medium risk: 0.30). Source text from outsider-authored web pages can reach the LLM via the website crawl stage that performs WebFetch of the user-derived website URLs and then extracts visible text from the fetched pages in references/business-context.md/Phase 4.

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

• Potentially malicious external URL detected (high risk: 0.80). The skill explicitly instructs using a runtime runScript that extends the server resource notfair://playbooks/audit-account, meaning the agent will fetch and execute/extend remote script content at runtime (notfair://playbooks/audit-account) which can directly control code and behavior.