The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses shell commands to facilitate a local development environment. It instructs the agent to run pnpm exec vite to host a preview gallery and uses kill $(lsof -ti tcp:<port>) to terminate the specific Vite process after testing. These are standard operations for a development and verification skill.
[DYNAMIC_EXECUTION]: The skill recommends generating a temporary driver script (node_modules/.drive.mjs) to execute Playwright tests. This runtime script generation and execution is a common pattern for automating browser-based end-to-end verification.
[INDIRECT_PROMPT_INJECTION]: The skill creates a surface for indirect injection by ingesting and rendering metadata designer configurations and sample drafts for testing purposes.
Ingestion points: Metadata designer files in packages/app-shell/src/views/metadata-admin/previews/* and sample data in apps/console/src/preview-samples.ts.
Boundary markers: No specific delimiters or warnings for embedded instructions are provided in the documentation.
Capability inventory: Includes starting network services (Vite), executing Node.js scripts (Playwright), and process management.
Sanitization: No specific sanitization or validation of the processed metadata is mentioned, as the tool is intended for internal developer verification.

verify