extract
Pass
Audited by Gen Agent Trust Hub on Mar 31, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill is entirely instruction-based and provides a methodology for design system maintenance. No malicious executable code or suspicious automation patterns were found.
- [PROMPT_INJECTION]: The use of markers like 'CRITICAL' and 'IMPORTANT' is restricted to task-specific guidance (e.g., asking before creating new directories) and does not attempt to override the AI agent's core safety instructions or bypass constraints.
- [DATA_EXFILTRATION]: There are no network calls, external URLs, or instructions to access sensitive configuration files like '.env' or SSH keys. The 'grep' operations suggested are limited to identifying UI-related keywords in the local project.
- [REMOTE_CODE_EXECUTION]: No external dependencies are fetched, and there is no dynamic execution of code from untrusted sources. All actions involve standard file-system operations for code organization.
Audit Metadata