frontend-design-ultimate

Pass

Audited by Gen Agent Trust Hub on Mar 13, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill downloads numerous dependencies and framework boilerplates from the NPM registry using standard CLI tools such as create-vite, create-next-app, and the shadcn CLI. These resources are from well-known and established services in the web development ecosystem.
  • [COMMAND_EXECUTION]: The skill utilizes shell scripts (init-vite.sh, init-nextjs.sh, bundle-artifact.sh) to automate project setup, dependency management, and production bundling. These scripts perform routine development tasks like writing configuration files and executing build commands.
  • [PROMPT_INJECTION]: The skill processes natural language design requirements from the user to generate production-ready code. This represents an indirect prompt injection surface where untrusted data (user requirements) is used to influence the agent's code output.
  • Ingestion points: User requirements for site design provided to the agent (referenced in SKILL.md).
  • Boundary markers: None present; user inputs are processed directly to determine aesthetic and functional requirements.
  • Capability inventory: The agent can execute shell scripts to install packages and write files to the local file system (found in scripts/).
  • Sanitization: None present; the skill relies on the agent's logic to interpret user prompts into code templates.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 13, 2026, 06:59 AM