postgres-pro

Pass

Audited by Gen Agent Trust Hub on Sep 23, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONPRIVILEGE_ESCALATION
Full Analysis
  • [COMMAND_EXECUTION]: The scripts/backup_pg.py script uses subprocess.run to execute system utilities such as pg_dump, psql, and pg_restore for database management tasks.
  • [INDIRECT_PROMPT_INJECTION]:
  • Ingestion points: User-provided database names and backup names are passed to the management methods in scripts/backup_pg.py.
  • Boundary markers: None present in the script logic.
  • Capability inventory: The script possesses the capability to execute shell commands and interact with the database engine via subprocess.run calls to psql.
  • Sanitization: The _create_database method in scripts/backup_pg.py uses string interpolation (f"CREATE DATABASE {database};") to build a command for psql -c. This does not sanitize the database variable, allowing a malicious user to potentially inject SQL fragments if the input is not validated by the agent.
  • [PRIVILEGE_ESCALATION]: Documentation in EXAMPLES.md and REFERENCE.md includes instructions for using sudo to perform necessary administrative tasks, such as modifying system configuration files (postgresql.conf) and managing the PostgreSQL service via systemctl.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 23, 2026, 08:05 AM
Security Audit — agent-trust-hub — postgres-pro