setup
Pass
Audited by Gen Agent Trust Hub on Oct 4, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes shell commands to scaffold a new project using 'npx create-next-app', initialize Git, and start/stop a development server for verification purposes.
- [EXTERNAL_DOWNLOADS]: The skill uses 'npx' to download the latest version of the Next.js scaffolding tool from the public npm registry, which is standard behavior for modern web development workflows.
- [INDIRECT_PROMPT_INJECTION]: The skill accepts user arguments to determine the technology stack and generates a file named 'AGENTS.md' that contains instructions governing the agent's future behavior, such as 'Make reasonable assumptions instead of asking clarifying questions'. These instructions are design choices to increase agent autonomy during the development process.
Audit Metadata