The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill installs numerous frontend development packages (including React 18, Vite, Tailwind CSS, Parcel, and shadcn/ui components) from the official NPM registry via pnpm and npm. These are standard, well-known development tools.
[COMMAND_EXECUTION]: Automated shell scripts (init-artifact.sh and bundle-artifact.sh) perform project setup, configuration editing using sed and node, and asset bundling. These operations are functional requirements for the skill's stated purpose of building web applications.
[INDIRECT_PROMPT_INJECTION]: As a generative tool for creating web artifacts, there is a theoretical surface for indirect prompt injection if untrusted user data is interpolated into the generated HTML/JSX code.
Ingestion points: User prompts and instructions used to define the features or content of the generated web artifact (processed in SKILL.md workflows).
Boundary markers: No specific boundary markers or "ignore" instructions are implemented in the setup scripts to delimit external data.
Capability inventory: File system write access (for source code and config), execution of shell-based build tools, and installation of Node.js packages.
Sanitization: The skill relies on the AI agent's generation phase for safety and does not include explicit sanitization or validation logic within the provided shell scripts.

artifacts-builder