The Agent Skills Directory

[SAFE]: The skill serves as a documentation repository for Nuxt UI v4, offering templates and best practices for building Vue-based user interfaces. It does not contain any executable scripts or malicious instructions.
[EXTERNAL_DOWNLOADS]: The documentation includes instructions for installing official packages such as @nuxt/ui, tailwindcss, and AI integration libraries like @ai-sdk/vue and @comark/nuxt using standard package managers (pnpm). These are legitimate, industry-standard dependencies from the vendor.
[COMMAND_EXECUTION]: The skill provides examples of standard development commands, such as using npx nuxi@latest init to bootstrap projects from official Nuxt templates and claude mcp add to configure the official Nuxt UI MCP server.
[INDIRECT_PROMPT_INJECTION]: While the skill provides components for handling user-generated content (e.g., chat interfaces and forms), it mitigates risks by recommending robust sanitization patterns:
Ingestion points: Untrusted data is handled through components like UChatMessages, UForm, and UInput described in references/layouts/chat.md and references/guidelines/forms.md.
Boundary markers: The skill recommends using the Comark component to render AI-generated markdown, which handles content rendering in a controlled manner.
Capability inventory: The skill focus is on UI presentation; no dangerous system capabilities or arbitrary code execution paths are exposed.
Sanitization: The skill strongly promotes the use of Standard Schema (Zod) for all form inputs to ensure data integrity and validation before processing.

nuxt-ui