The Agent Skills Directory

[SAFE]: The skill provides a knowledge base of shell scripting best practices. While it contains examples of dangerous commands (e.g., 'rm -rf /*'), these are explicitly used within 'Incorrect' code blocks to illustrate vulnerabilities and the necessity of safety measures like 'set -u'.\n- [SAFE]: External references point to authoritative and trusted sources such as the Google Shell Style Guide, ShellCheck documentation, and the POSIX specification. No unverified third-party dependencies or remote scripts are used or recommended.\n- [SAFE]: The skill contains no obfuscation, credential leaks, or exfiltration patterns. It promotes security-centric coding patterns like 'eval' avoidance, input validation, and secure temporary file creation.\n- [SAFE]: No persistence mechanisms or privilege escalation attempts were found. Examples involving 'sudo' or 'chmod' are illustrative of standard administrative tasks and security controls.\n- [SAFE]: The skill does not contain any dynamic context injection or meta-instruction patterns meant to bypass agent safety filters.

shell