The Agent Skills Directory

Unverifiable Dependencies & Remote Code Execution (SAFE): The scripts init-artifact.sh and bundle-artifact.sh install many packages from the npm registry (e.g., vite, tailwindcss, parcel, radix-ui). These are well-known libraries from a trusted registry, and the downloads are essential for the skill's primary purpose.
Dynamic Execution (LOW): The initialization script uses node -e to programmatically modify tsconfig.json and tsconfig.app.json. While these are configuration changes, the use of dynamic string execution for file modification is noted.
Indirect Prompt Injection (LOW): This skill is a code generation engine, which creates a potential surface for indirect injection if malicious instructions are embedded in the user's design requests. \n
Ingestion points: User instructions provided to the agent are used to generate React source code files. \n
Boundary markers: No explicit delimiters are used in the build scripts to separate user-originated content from the build logic. \n
Capability inventory: The skill executes shell scripts, installs software via pnpm, and runs a build pipeline using parcel. \n
Sanitization: The provided scripts do not validate the generated code; security relies on the agent's internal safety filters and the runtime sandboxing of the resulting HTML artifact.

artifacts-builder