create-next-project

Pass

Audited by Gen Agent Trust Hub on Oct 7, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill uses the AskUserQuestion tool in SKILL.md (Step 2) to gather project requirements from the user, which are then used to direct the behavior of specialized build agents.
  • Ingestion points: Project requirements, theme presets, and optional package selections gathered via AskUserQuestion in SKILL.md.
  • Boundary markers: None identified in the instructions to separate user input from agent instructions.
  • Capability inventory: The build agents dispatched in Step 4 have capabilities to write files and execute shell commands (bun, git, gh, vercel, convex).
  • Sanitization: No explicit sanitization or validation of user-supplied strings is described before they are interpolated into project files or used in configuration.
  • [COMMAND_EXECUTION]: The skill utilizes several standard command-line interfaces to perform its primary purpose of project scaffolding and deployment.
  • Evidence: Instructions in SKILL.md and references/stack-defaults.md invoke bun, git, gh, vercel, and convex to initialize repositories, install dependencies, and provision infrastructure.
  • [EXTERNAL_DOWNLOADS]: The skill installs numerous third-party dependencies from public registries.
  • Evidence: The skill automates the installation of packages such as better-auth, convex, @tanstack/react-query, and others via the bun add command. These are standard packages used for the skill's stated purpose.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 7, 2026, 05:28 PM