postgres-advanced

Pass

Audited by Gen Agent Trust Hub on Apr 8, 2026

Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill utilizes subprocess.run to call the pg_dump utility for creating database backups. It adheres to security best practices by passing arguments as a list to prevent shell injection and enforcing a five-minute timeout. (Evidence: pg_tool/services/backup.py)
  • [COMMAND_EXECUTION]: The skill executes SQL queries provided by the agent via the pg query command. To prevent unauthorized or destructive operations, it implements a heuristic safety scanner that blocks commands like DROP, ALTER, and TRUNCATE, as well as multi-statement execution. (Evidence: pg_tool/services/safety.py, pg_tool/services/query.py)
  • [SAFE]: Database schema migrations are loaded dynamically using importlib.import_module and pkgutil.iter_modules. This implementation is restricted to local modules within the specific pg_tool.migrations package, following standard patterns for plugin or migration management. (Evidence: pg_tool/migrations/__init__.py)
  • [SAFE]: Database connection credentials (DSN) are managed via the PG_DSN environment variable. The skill documentation and implementation provide clear guidance on avoiding hardcoded secrets and leveraging environment-based configuration for secure state management. (Evidence: pg_tool/config.py, SKILL.md)
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 8, 2026, 02:44 PM
Security Audit — agent-trust-hub — postgres-advanced