The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructions and documentation frequently reference the use of the shadcn CLI via npx for project initialization, component installation, and migrations (e.g., npx shadcn@latest init, npx shadcn@latest add, npx shadcn@latest migrate rtl). Additionally, the skill includes a utility script scripts/verify-setup.sh that performs local filesystem checks and package dependency validation.
[INDIRECT_PROMPT_INJECTION]: The skill has an indirect prompt injection surface because it is designed to ingest and process external, potentially untrusted content (component source code, metadata, and blocks) from registries via MCP tools, which is then written to the local project directory.
Ingestion points: Data enters the agent context through MCP tools such as get_component, get_block, get_component_metadata, and search_items_in_registries (referenced in SKILL.md and README.md).
Boundary markers: No specific delimiters or "ignore embedded instructions" warnings are used when processing the fetched component code.
Capability inventory: The skill possesses Bash, Write, and Read capabilities, allowing it to execute scripts and modify the local filesystem based on ingested data.
Sanitization: The skill does not explicitly describe sanitization or validation of the remote component code before writing it to the filesystem.
[EXTERNAL_DOWNLOADS]: The skill facilitates the download and installation of numerous standard React UI components and library dependencies from the npm registry. It also references official documentation and source code from well-known services including Radix UI, Tailwind CSS, and GitHub repositories associated with the shadcn/ui ecosystem.

shadcn-ui