claude-design-system-prep

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

• Third-party content exposure detected (high risk: 0.85). Required workflow includes WebFetch로 운영 중 웹사이트 URL의 라이브 페이지를 분석(2단계 자산 분석)하여 페이지의 자유형 텍스트/콘텐츠를 LLM 컨텍스트에 읽히게 하는 경로가 존재함(OUTSIDER: public web content fetched at runtime).

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

• Potentially malicious external URL detected (high risk: 0.90). The skill explicitly uses the external "moai-design:design-system-library" (e.g., systems/registry.md and systems/.md) at runtime as the primary source whose YAML token content is injected directly into the generated DESIGN.md, so that fetched content directly controls the agent's output.