database

Pass

Audited by Gen Agent Trust Hub on Mar 4, 2026

Risk Level: SAFEREMOTE_CODE_EXECUTIONCOMMAND_EXECUTION
Full Analysis
  • [REMOTE_CODE_EXECUTION]: The skill dynamically imports and executes JavaScript files from the src/db/migrations/ and src/db/seeds/ directories. While this is the intended mechanism for running database migrations and seeds, it creates a vector for executing arbitrary code if malicious files are placed in these directories.
  • Ingestion points: The skill reads file names from the local file system using readdir in src/db/migrate.js and src/db/seed.js.
  • Boundary markers: No boundary markers or integrity checks are present to verify the source or content of the scripts before execution.
  • Capability inventory: Loaded scripts execute arbitrary logic and database queries via the provided db pool.
  • Sanitization: No validation or sandboxing of the script content is performed.
  • [COMMAND_EXECUTION]: The skill is designed to execute arbitrary SQL queries defined in migration and seed files. This functionality is essential for database management but allows for high-impact operations that must be restricted to trusted migration scripts.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 4, 2026, 06:11 AM