The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes shell commands to manage the lifecycle of a design preview environment.
It uses node scripts/design-preview.mjs to launch a local server and kill or pkill to terminate the background process once the user selection is complete.
[REMOTE_CODE_EXECUTION]: The skill relies on executing a local script (scripts/design-preview.mjs) within the project directory to provide its core functionality. This script implements the HTTP server and API logic for the design selection process.
[PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface (Category 8) due to its data-ingestion and generation logic.
Ingestion points: The skill reads user-supplied design descriptions and alternatives from the 02-design.md file located in the feature directory.
Boundary markers: No explicit boundary markers or "ignore embedded instructions" warnings are documented for the HTML generation process.
Capability inventory: The skill can write files (design-comparison.html, design-selection.json, 02-design.md), start a local HTTP server, and execute shell commands to manage server processes.
Sanitization: The instructions do not specify any HTML sanitization or escaping protocols for the content extracted from the markdown file before it is interpolated into the generated HTML preview page.

design-interactive-preview