The Agent Skills Directory

[COMMAND_EXECUTION]: The skill documentation and README instruct the execution of a local PowerShell script (scripts/search.ps1) using the -ExecutionPolicy Bypass flag. This flag is used to circumvent local script execution restrictions, which is a security-weakening practice, although common in automation contexts.
[PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface by processing untrusted user input to generate executable files.
Ingestion points: Product/domain descriptions and page types provided by the user at runtime.
Boundary markers: None identified in the workflow to separate user instructions from design data.
Capability inventory: The skill is capable of generating and writing files (specifically index.html and styles.css) to the filesystem, as evidenced in the SKILL.md workflow.
Sanitization: There is no evidence of input validation or escaping mechanisms to prevent user-supplied text from being maliciously interpolated into the generated HTML/CSS files.
[COMMAND_EXECUTION]: The skill references a script file scripts/search.ps1 which is not included in the provided file list. The absence of this script prevents a complete security audit of its internal logic, such as whether it performs network operations or handles files securely.

ui-design-optimizer