scaffold-nextjs

Pass

Audited by Gen Agent Trust Hub on Mar 13, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: Configures the shadcn CLI to fetch UI components from the vendor's registry at https://ui.blode.co/r/{name}.json during Phase 3.
  • [EXTERNAL_DOWNLOADS]: Installs several NPM packages to provide project functionality, including agentation, ultracite, blode-icons-react, and turbo.
  • [COMMAND_EXECUTION]: Executes a series of system commands using npx, npm, gh, and vercel to automate project creation, workspace management, and deployment.
  • [PROMPT_INJECTION]: Contains an indirect prompt injection surface (Category 8) by interpolating user-provided inputs such as {{name}} and {{repo}} directly into shell commands (npx create-next-app, gh repo create) and project configuration files without explicit sanitization or boundary markers.
  • Ingestion points: Project metadata gathered in Phase 1 (SKILL.md).
  • Boundary markers: Absent.
  • Capability inventory: Subprocess execution for project setup, dependency installation, and git/deployment operations (references/app-setup.md, references/deploy-and-launch.md).
  • Sanitization: No input validation or escaping logic is defined in the instructions.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 13, 2026, 09:58 AM