The Agent Skills Directory

[SAFE]: The skill uses the Deno runtime to perform local file system operations. All scripts require explicit user-granted permissions (--allow-read and --allow-write) and do not perform any unauthorized actions outside the specified scope of design auditing and asset generation.
[SAFE]: No network operations or external downloads were detected. The scripts use standard Deno library imports via JSR (e.g., @std/cli, @std/path, @std/fs), which are trusted modules for the Deno ecosystem.
[SAFE]: The code generation logic in generate-component.ts and other generation scripts uses static templates and basic string interpolation. No dynamic code execution (e.g., eval or exec) of user-controlled input is present, ensuring that generated components are static templates.
[SAFE]: Analysis of the instructions and documentation shows no attempts at prompt injection or safety bypass. The provided design philosophy is strictly focused on aesthetic steering to improve the quality of UI output without manipulating the agent's core safety instructions.
[SAFE]: There are no hardcoded credentials, API keys, or exfiltration patterns. The skill processes local CSS and HTML files as intended for its primary purpose of design auditing.

frontend-design