brand-onboarding

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

• Third-party content exposure detected (medium risk: 0.30). In “Brand Guidelines Onboarding,” outsiders can supply free-text brand content via “Paste brand document text” (Path A) or “Provide file path or URL to brand document” (Path B), which the wizard then extracts and ingests at runtime to generate brand-guidelines.md.

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

• Potentially malicious external URL detected (high risk: 0.80). The skill explicitly fetches user-supplied URLs and downloads logo files at runtime (e.g., https://acme.com/brand, https://acme.com/brand-style-guide, https://company.com/assets/logo.svg, https://company.com/logo.png), and the fetched content is extracted and injected into the agent's context to drive prompt behavior and guideline generation.