brand-intel-branddev

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

• Third-party content exposure detected (high risk: 0.85). The skill’s runtime workflow calls Brand.dev endpoints that crawl/scrape arbitrary user-supplied domains (e.g., /v1/brand/retrieve, /v1/brand/fonts, /v1/brand/styleguide, /v1/brand/screenshot, /v1/brand/ai/query), so outsider-authored website text is ingested into the LLM context via the website-crawling/AI-extraction path.