The Agent Skills Directory

[SAFE]: The skill consists of Markdown-based design guidelines and instructions for the agent to perform visual audits. It does not include any executable scripts or binary files.
[SAFE]: Tool usage is restricted to Read, Glob, and Grep in the allowed-tools frontmatter, which are appropriate for analyzing source code for UI improvements.
[SAFE]: No suspicious command execution (e.g., sudo, chmod), network requests to unknown domains, or sensitive file access patterns (e.g., .ssh, .aws) were detected.
[SAFE]: External links provided in the documentation are restricted to legitimate design reference sites and standard affiliate links for the referenced design books. These do not pose a security risk to the agent or user environment.
[SAFE]: Although the skill processes untrusted user data (project source code), its capabilities are limited to reading and reporting, with no write or network permissions that could be leveraged for malicious data exfiltration.

refactoring-ui