vue-layout

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

• Third-party content exposure detected (high risk: 0.75). Runtime path: Step 1/2 uses “Figma URL” and Framelink MCP for Figma to read design nodes; this ingests outsider-authored Figma content (public/third-party design artifacts) into the agent’s LLM context for interpretation.

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

• Potentially malicious external URL detected (high risk: 1.00). The skill requires a Figma URL (https://www.figma.com/file/{fileId}/...?node-id={nodeId}) which at runtime is read via the Framelink MCP (e.g., get_node(fileId, nodeId)) and the fetched design nodes/styles directly drive the agent's instructions/output, making this a required external runtime dependency.