The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides defensive guidelines for executing native programs and PowerShell cmdlets. It explicitly discourages dangerous practices like Invoke-Expression and shell command string concatenation, instead recommending structured argument passing via arrays and ProcessStartInfo to prevent command injection.
[DYNAMIC_EXECUTION]: The skill recommends writing complex commands to a temporary .ps1 file to avoid quoting issues, which is a secure method for handling complex logic compared to raw string execution. This is identified as a low-risk, safe practice for template-based script generation.
[SAFE]: No malicious patterns, remote code execution, or unauthorized data access were found in the provided files. The instructions promote secure patterns that actively mitigate common security risks like path traversal and shell injection.

powershell-safe-invocation