create-property

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

• Third-party content exposure detected (high risk: 0.80). The skill explicitly fetches and ingests user-provided Figma file content (e.g., Step 4 "Extract Property Data" via figma_execute reading node names, componentPropertyDefinitions, slot descriptions; Step 4b calling figma_get_variables to read variable collection names/IDs) and then uses that extracted, user-generated text in the AI Validation / Exhibit Planning (Step 4e) to decide exhibit plans and rendering behavior, so arbitrary third‑party content in a Figma file could influence the agent's decisions.