create-anatomy

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

• Third-party content exposure detected (high risk: 0.80). The skill explicitly ingests arbitrary user-supplied Figma files/URLs (Step 3: "Navigate to the component file and run the extraction script" and code that walks figma.root.children and reads node text/names/children), treats that untrusted, user-generated content as input to AI reasoning (Step 4 enriches notes and may re-extract based on the extracted data) and uses it to drive actions (unhiding, slot population, creation of sections), so third-party content can materially influence agent behavior.