atomic-design

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

• Third-party content exposure detected (high risk: 0.90). This skill mandates extracting and interpreting design data from external Figma files via the Figma MCP (e.g., figma.get_node, figma.get_local_variables, and user-provided Figma links / MCP server), meaning the agent will fetch and process arbitrary third-party/user-generated content that could carry indirect prompt injection.

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

• Potentially malicious external URL detected (high risk: 0.80). The skill mandates runtime extraction from Figma (e.g., fetching a Figma file link like https://figma.com/file/xxx/Design-System?node-id=123 and calling the remote MCP endpoint https://mcp.figma.com/mcp) so external design content would be fetched at runtime and directly drive the agent's implementation instructions, creating a high-confidence external dependency that can control prompts.