linux-shell-scripting

Warn

Audited by Gen Agent Trust Hub on Jun 13, 2026

Risk Level: MEDIUMCREDENTIALS_UNSAFECOMMAND_EXECUTIONREMOTE_CODE_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
  • [CREDENTIALS_UNSAFE]: The database backup script defines a password variable and passes it to the mysqldump command as a CLI argument (-p"$db_pass"). This is an insecure practice as passwords can be exposed in process lists to other users on the system.\n- [COMMAND_EXECUTION]: Multiple scripts utilize sudo to perform high-privilege operations such as package installation and service management. Additionally, the task scheduler script modifies the user's crontab, establishing a persistence mechanism for recurring code execution.\n- [REMOTE_CODE_EXECUTION]: The skill includes a template for remote script execution that pipes local script contents to a remote bash shell over SSH (ssh ... "bash -s" < ...).\n- [DATA_EXFILTRATION]: Remote backup and synchronization templates allow for the transfer of local data to external remote servers using rsync and ssh.\n- [PROMPT_INJECTION]: The skill processes untrusted system and web server logs, creating an indirect prompt injection surface.\n
  • Ingestion points: /var/log/syslog and /var/log/apache2/access.log (referenced in SKILL.md).\n
  • Boundary markers: No delimiters or instructions are used to isolate untrusted log data from the processing logic.\n
  • Capability inventory: The skill has capabilities for high-privilege command execution (sudo), file deletion (rm), and network communication (ssh, rsync, curl).\n
  • Sanitization: No validation or escaping is applied to log content before it is processed by shell utilities like grep and awk.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Jun 13, 2026, 02:07 AM
Security Audit — agent-trust-hub — linux-shell-scripting