ux-foundation

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

• Third-party content exposure detected (medium risk: 0.30). In references/figma-integration.md, when Figma mockups are enabled it can read Figma node metadata/tokens and generate design content via the Figma MCP (get_metadata, get_variable_defs, etc.), and those inputs can be influenced by an outsider who writes content into the shared Figma file, making free-text ingestion possible at runtime.

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

• Potentially malicious external URL detected (high risk: 0.80). The Figma integration explicitly directs the agent at runtime to load external MCP guidance resources via "skill://figma//SKILL.md" (and to use Figma file links like "figma.com/design/:fileKey/..."), which will be fetched/loaded during skill execution and can change how the agent calls tools or behaves.