nextjs-project-starter
Pass
Audited by Gen Agent Trust Hub on Mar 6, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes shell commands for project creation and configuration, including 'npx create-next-app', 'npm install', 'gh repo create', and 'vercel link'. These commands use parameters derived from user input.
- [EXTERNAL_DOWNLOADS]: The skill fetches package information from the npm registry and installs dependencies from official repositories. It references well-known services including GitHub, Vercel, and Supabase, which are considered safe and trusted sources.
- [PROMPT_INJECTION]: The skill has an indirect prompt injection surface due to user input interpolation into system commands.
- Ingestion points: 'Project name' and 'Project description' inputs in SKILL.md.
- Boundary markers: Absent; user input is directly concatenated into shell commands.
- Capability inventory: Subprocess execution ('npx', 'npm', 'gh', 'vercel') and file writing ('echo') as described in Phase 6 of the workflow.
- Sanitization: Absent; instructions request kebab-case for the project name but do not provide programmatic enforcement or escaping in the workflow instructions.
Audit Metadata