design-system

Pass

Audited by Gen Agent Trust Hub on Jul 29, 2026

Risk Level: SAFE
Full Analysis
  • External Data Integration: The skill fetches design tokens and assets from well-known services like Figma and user-provided URLs. These operations are governed by a strict security policy that enforces HTTPS, limits redirect chains, and prevents access to private network addresses to mitigate server-side request forgery (SSRF).
  • Local File Security: The skill interacts with local project files, including archives and codebases, to extract brand information. It implements rigorous path-safety checks to prevent traversal attacks and blocks access to sensitive system directories and credential stores (e.g., SSH and AWS configurations).
  • Input Sanitization and Injection Defense: Instructions explicitly include a security layer that sanitizes untrusted user content from files and URLs. This includes stripping potential prompt injection patterns and wrapping external data in distinct delimiters before processing by the model.
  • Controlled Command Execution: The skill uses shell commands for project organization, such as creating snapshots and opening local preview files. These are limited to benign administrative tasks within the project's working directory.
  • Deterministic Generation: The generation of design artifacts like TypeScript tokens and HTML galleries is performed using fixed schemas and templates, ensuring consistent and predictable output without runtime code execution risks.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 29, 2026, 03:31 AM
Security Audit — agent-trust-hub — design-system