figma-design-system-to-design-md

Warn

Audited by Gen Agent Trust Hub on Jul 30, 2026

Risk Level: MEDIUMREMOTE_CODE_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [REMOTE_CODE_EXECUTION]: The skill provides installation instructions that involve downloading and running a plugin from an unverified GitHub repository (https://github.com/albertzhangz10/figma-design-system-to-design-md). This repository does not belong to a recognized trusted organization and is hosted on a personal account that does not match the vendor's claimed domain.
  • [EXTERNAL_DOWNLOADS]: The skill references multiple external web resources, including its own web interface (https://figmadesignmd.com/) and design system documentation hosted by Google (https://stitch.withgoogle.com/docs/design-md/overview).
  • [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection due to its core function of ingesting and parsing data from potentially untrusted files.
  • Ingestion points: It automatically searches for and reads CSS, JSON, JavaScript, and TypeScript token files, as well as Tailwind and Uno configuration files.
  • Boundary markers: There are no defined delimiters or instructions to prevent the model from executing instructions that might be embedded within the design token values or configuration files.
  • Capability inventory: The skill reads the local filesystem, executes plugin installation commands, and integrates with the Figma MCP server to fetch external design data.
  • Sanitization: The instructions do not implement any mechanisms to escape, validate, or filter content retrieved from the detected project files or external Figma API responses.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Jul 30, 2026, 01:39 AM
Security Audit — agent-trust-hub — figma-design-system-to-design-md