The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes dynamic context injection in SKILL.md to execute npx shadcn@latest info --json. This command runs automatically upon loading the skill to gather project-specific metadata (such as the detected framework, Tailwind version, and import aliases), which is then used to ensure the agent's generated code is compatible with the user's environment.
[EXTERNAL_DOWNLOADS]: The skill facilitates the download of UI components and documentation from official and well-known registries. It uses the shadcn CLI to pull component source code and suggests fetching documentation and example implementations from ui.shadcn.com and GitHub repositories. These sources are considered reputable and are intended for the primary purpose of the skill.
[PROMPT_INJECTION]: The skill processes external data from documentation URLs and registry items, which constitutes an indirect prompt injection surface. This is mitigated by explicit instructions for the agent to use --dry-run, --view, and --diff flags to inspect external code before installation, alongside a mandatory workflow requiring the agent to read and verify all newly added files for correctness and design compliance.

shadcn