bash-defensive-patterns

Pass

Audited by Gen Agent Trust Hub on Jul 19, 2026

Risk Level: SAFE
Full Analysis
  • [PROMPT_INJECTION]: No evidence of malicious instructions, safety bypasses, or adversarial role-play prompts were detected. The skill provides legitimate educational content for shell scripting.
  • [DATA_EXFILTRATION]: No commands for exfiltrating sensitive data or hardcoded credentials were found. The skill actively encourages safe variable handling and input validation.
  • [REMOTE_CODE_EXECUTION]: The skill does not perform or encourage remote code execution from untrusted sources. It references documentation from trusted and well-known technical resources for educational purposes.
  • [COMMAND_EXECUTION]: All included code templates utilize standard system commands (e.g., mv, rm, mkdir) in a safe and defensive manner, emphasizing variable quoting and error trapping to prevent command injection or unintended file operations.
  • [CREDENTIALS_UNSAFE]: No hardcoded API keys, tokens, or private keys were found. The templates use placeholders correctly or focus on environment variable validation.
  • [SAFE]: The skill explicitly advises against running scripts as root and promotes the use of dry-run flags for destructive operations, which aligns with security best practices for automation.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 19, 2026, 04:13 AM
Security Audit — agent-trust-hub — bash-defensive-patterns