where-your-customer-lives

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

• Third-party content exposure detected (medium risk: 0.30). The script fetches public Reddit API responses, Hacker News search results, and DuckDuckGo search snippets and web pages, which contain outsider-authored free text ingested into the runtime.

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

• Potentially malicious external URL detected (medium risk: 0.30). The skill makes programmatic HTTP calls to the live Reddit and Hacker News APIs and scrapes DuckDuckGo HTML at runtime to fetch external data and search results, constituting common/reputable unofficial inbound operational dependencies.