The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes shell commands like grep, wc, and python3 -m http.server to audit source code and host a local environment for visual inspection. These commands are integral to the skill's functionality as a design quality gate.
[DYNAMIC_EXECUTION]: The skill executes JavaScript within a browser environment and uses node -e to calculate layout properties and text legibility, which is a standard procedure for UI validation during the auditing process.
[INDIRECT_PROMPT_INJECTION]: The skill processes project artifacts (HTML, CSS, Markdown) which could contain malicious instructions. The risk is mitigated by the skill's design, which mandates isolation between analysis phases and requires a blind 'challenger' gate for final approval.

design-review