The Agent Skills Directory

[Dynamic Script Execution]: The skill generates JavaScript files (".cjs") based on issue descriptions and executes them using "node" to automate VS Code UI interactions. While this is the core functionality required for testing, executing dynamically generated code is a pattern that warrants standard developer review.
[Environment Configuration with Elevated Privileges]: The documentation suggests using "sudo apt install" to install the "ffmpeg" dependency on Linux systems. This procedure involves elevated privileges to prepare the execution environment.
[External Content Processing]: The skill derives automation logic from external GitHub issue descriptions. This creates a surface for indirect prompt injection, where content within an issue could influence the generated test scripts.
[External Package and Tool Installation]: The skill instructs the agent to perform "npm install" and use platform package managers (winget, brew, apt) to fetch dependencies. These operations involve downloading and installing software from external registries.
[File System Access]: The skill creates temporary directories using "os.tmpdir()" and writes evidence artifacts, such as videos and reports, to the ".build" folder within the project directory. These operations are scoped to the intended use case of generating test evidence.

validate-ui-scenario