The Agent Skills Directory

[COMMAND_EXECUTION]: The script scripts/init-artifact.sh passes the user-provided project-name argument directly into a sed command and other shell operations without sanitization. This allows for potential command injection or file system manipulation if a project name contains shell metacharacters or characters that break the sed delimiter.
[PRIVILEGE_ESCALATION]: The scripts/init-artifact.sh script includes a command to install the pnpm package manager globally (npm install -g pnpm) if it is not found on the system. Modifying global system packages is a privileged operation that changes the execution environment for other users or processes.
[INDIRECT_PROMPT_INJECTION]: The skill builds React-based HTML artifacts from user-driven development. It lacks input validation or sanitization for content that the AI agent might write into the artifact files, which could lead to the creation of artifacts containing malicious scripts (XSS) if the agent is misled by external data during development. 1. Ingestion points: The project name argument in scripts/init-artifact.sh and the source code generated during the development phase. 2. Boundary markers: None present. 3. Capability inventory: The skill has the ability to write to the file system, install npm packages, and execute build processes using Parcel and Vite. 4. Sanitization: There is no evidence of sanitization or escaping of user-influenced strings before they are written to configuration files or bundled into the final HTML.

web-artifacts-builder