The Agent Skills Directory

[COMMAND_EXECUTION]: The scripts/init-artifact.sh script is vulnerable to command injection via the project name argument. The variable $PROJECT_NAME is interpolated into a sed command using unsafe shell concatenation ('s/.../...'"$PROJECT_NAME"'/'). An attacker providing a project name like test' ; touch RCE ; # could execute arbitrary shell commands on the host system.
[EXTERNAL_DOWNLOADS]: The skill performs extensive dynamic package installations at runtime using npm and pnpm. While these target the official registries, the sheer volume of dependencies (Vite, Tailwind, Parcel, Radix UI, etc.) increases the potential attack surface for supply chain compromises.
[COMMAND_EXECUTION]: The script scripts/init-artifact.sh installs the pnpm package manager globally (npm install -g pnpm) if it is not found. Global installations typically require elevated permissions and modify the system state outside of the project directory.
[EXTERNAL_DOWNLOADS]: The skill relies on a local archive shadcn-components.tar.gz which is extracted during initialization. As the contents of this tarball are not provided in the source for analysis, it represents an opaque binary blob that could contain malicious scripts or unauthorized code which would be executed or included in the final artifact.

artifacts-builder