ads-dna

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

• Third-party content exposure detected (high risk: 0.95). The skill uses a browser/curl tool at runtime to fetch arbitrary user-supplied URLs and then extracts “all visible text content” plus meta/style/og:image values from those fetched pages, which is outsider-authored free text from public web content.

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

• Potentially malicious external URL detected (high risk: 0.80). The skill explicitly fetches arbitrary, user-provided website URLs at runtime (e.g., https://yoursite.com) via the browser/curl tool and injects the extracted page content into the generated brand-profile.json and downstream prompts, meaning external site content directly controls the agent's outputs.