cra-to-vite-migrator
Pass
Audited by Gen Agent Trust Hub on Aug 3, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [SAFE]: The skill implements a robust three-phase migration workflow (Inspect, Plan, Execute) that mandates a manual review and user approval before any file modifications occur.\n- [SAFE]: Explicit security rules are defined in SKILL.md, prohibiting the use of
curl | bash,eval, and file operations outside the project directory, while requiring backups before changes.\n- [COMMAND_EXECUTION]: The skill uses standard package management commands (npm install,npm uninstall) and a local scanning script. The scriptscripts/migrate-env.shcontains security hardening, includingset -euo pipefailand path traversal validation (/../check).\n- [EXTERNAL_DOWNLOADS]: The skill manages the installation of well-known, legitimate developer tools and plugins from the official Node.js registry (e.g.,vite,vitest,@svgr/rollup).\n- [PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection (Category 8) as it reads and processes untrusted project source code. \n - Ingestion points: Project source files (JS/TS), configuration files (package.json, jsconfig.json), and environment files (.env) are read during Phase 1.\n
- Boundary markers: The agent is instructed to summarize risks and wait for user approval of a written plan (Phase 2) before executing changes.\n
- Capability inventory: Capabilities include file system writes,
npm install, and execution of a local bash helper script.\n - Sanitization: The skill requires JSON validation for
package.jsonand syntax validation forvite.config.jsbefore writing. The helper script uses safe argument handling to prevent shell injection.
Audit Metadata