linux-shell-scripting

Fail

Audited by Gen Agent Trust Hub on Jun 18, 2026

Risk Level: HIGHCREDENTIALS_UNSAFECOMMAND_EXECUTIONDATA_EXFILTRATIONREMOTE_CODE_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [CREDENTIALS_UNSAFE]: The Database Backup Script contains hardcoded placeholder credentials within the script body.
  • Evidence: db_user="username" and db_pass="password" defined in the Phase 1 Database Backup Script section.
  • [COMMAND_EXECUTION]: The skill utilizes powerful system commands and requests elevated privileges to perform administration tasks.
  • Evidence: Use of sudo for automated package installation via apt-get and service management via systemctl.
  • Evidence: Execution of administrative commands for account management including useradd, passwd, and chage.
  • [COMMAND_EXECUTION]: The skill implements persistence mechanisms on the host system.
  • Evidence: The Task Scheduler script programmatically modifies the user's crontab to ensure scripts execute automatically at recurring intervals.
  • [DATA_EXFILTRATION]: The skill includes scripts designed to move local data to remote network locations.
  • Evidence: The Remote Server Backup script utilizes rsync to transfer directory contents to an external server.
  • [REMOTE_CODE_EXECUTION]: The skill provides functionality to run shell commands on external infrastructure.
  • Evidence: The Remote Script Execution script uses ssh to pipe local script content to a remote server's bash shell.
  • [EXTERNAL_DOWNLOADS]: The skill verifies network status using well-known services.
  • Evidence: Performs connectivity checks and uptime monitoring targeting google.com and github.com using curl and ping.
Recommendations
  • AI detected serious security threats
Audit Metadata
Risk Level
HIGH
Analyzed
Jun 18, 2026, 09:14 PM
Security Audit — agent-trust-hub — linux-shell-scripting