The Agent Skills Directory

[DYNAMIC_CONTEXT_INJECTION]: The skill uses the ! command syntax in SKILL.md to execute npx shadcn@latest info --json when the skill is loaded. This is a benign use of project-specific tooling to provide the agent with necessary context about the project's configuration (framework, Tailwind version, aliases).
[EXTERNAL_DOWNLOADS]: The skill facilitates the installation of UI components from registries and GitHub repositories using the official shadcn CLI. It explicitly instructs the user to use the CLI rather than manual downloads, which ensures registry resolution and CSS diffing are handled correctly. Sources include well-known services like ui.shadcn.com and github.com.
[INDIRECT_PROMPT_INJECTION]: The skill provides an attack surface for indirect prompt injection by instructing the agent to fetch content from documentation and example URLs retrieved via the CLI.
Ingestion points: External URLs fetched via npx shadcn@latest docs from ui.shadcn.com and GitHub repositories.
Boundary markers: The skill does not define specific boundary markers or warnings for the content fetched from these external URLs.
Capability inventory: The skill has access to shell commands via npx shadcn@latest and file system operations to manage project components.
Sanitization: There is no mention of sanitizing or escaping the content fetched from external documentation or examples before the agent processes it.

shadcn