frontend-slides

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

• Third-party content exposure detected (high risk: 0.80). The SKILL.md explicitly accepts an arbitrary client website URL in "Use client branding" (Phase 2) and runs the visual-brand-extractor to read that site's visual identity and apply its CSS/variables, meaning the agent will fetch and interpret public third‑party web content as part of its workflow.