The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to execute shell commands to bootstrap the project environment, including pnpm create next-app, gh repo create, and vercel. These commands interpolate user-provided project names and repository identifiers.\n- [EXTERNAL_DOWNLOADS]: The skill fetches development tools and libraries from the npm registry, including agentation, blode-icons-react, and the author's own utility ultracite. It also configures the shadcn CLI to use a custom component registry hosted at blode.co.\n- [REMOTE_CODE_EXECUTION]: The skill utilizes pnpm dlx to download and run the latest versions of tools such as shadcn, ultracite, vercel, and knip at runtime.\n- [INDIRECT_PROMPT_INJECTION]: The skill generates AGENTS.md and CLAUDE.md files that serve as persistent instruction sets for AI agents working within the repository.\n
Ingestion points: User-provided project metadata ({{name}}, {{description}}) is interpolated into these agent-facing instruction files.\n
Boundary markers: The skill does not provide specific delimiters or warnings to prevent the agent from executing instructions potentially hidden within the user-supplied metadata.\n
Capability inventory: The scaffolded project environment provides the agent with extensive capabilities, including shell execution, filesystem modification, and network-based deployment tools.\n
Sanitization: There is no specified validation or sanitization of user-provided variables before they are written into the instruction files.

scaffold-nextjs