mongodb-crud

Pass

Audited by Gen Agent Trust Hub on Jul 15, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill interfaces with MongoDB using bash, ssh, and mongosh.\n
  • Evidence: The script mongodb_crud.sh uses standard system utilities for database operations.\n
  • Evidence: A static analysis hit for a reverse shell was found to be a false positive caused by /dev/tcp port scanning in find_free_port().\n- [SAFE]: Credentials and connection strings are managed with a focus on local security.\n
  • Evidence: Profiles are saved with 0600 permissions in ~/.config/mongodb-crud/profiles/.\n
  • Evidence: URI redaction is implemented in the redact_uri() function to prevent secret exposure in logs.\n- [SAFE]: Remote operations are conducted through documented SSH functionality.\n
  • Evidence: The ssh-remote mode executes a heredoc on a remote host to interface with MongoDB.\n
  • Evidence: This execution is gated by the same safety checks (read-only mode, dry-run) as local operations.\n- [SAFE]: Dynamic loading of configuration is performed using the Bash source command.\n
  • Evidence: The profile_path() function uses a safe_name regex (^[A-Za-z0-9_.-]+$) to prevent directory traversal attacks during sourcing.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 15, 2026, 03:40 PM
Security Audit — agent-trust-hub — mongodb-crud