The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides instructions for escalating privileges to Administrator level on Windows systems using the 'Start-Process -Verb RunAs' command to modify system settings.
[COMMAND_EXECUTION]: It enables the modification of critical system persistence vectors, including Windows Registry 'Run' keys (HKLM and HKCU), Scheduled Tasks, and System Services.
[REMOTE_CODE_EXECUTION]: The skill uses a pattern of generating PowerShell scripts, writing them to temporary files (e.g., /tmp/script.ps1), and executing them with bypassed execution policies via the '-ExecutionPolicy Bypass' flag.
[CREDENTIALS_UNSAFE]: The auditing logic captures full process command lines via WMI (Win32_Process), which often contains hardcoded credentials, session tokens, or sensitive environment variables passed as arguments at runtime.
[EXTERNAL_DOWNLOADS]: The skill references the author's public GitHub repository for additional resources and system maintenance context.
[PROMPT_INJECTION]: The skill processes untrusted system metadata (such as program names and process titles) and uses them in management commands without proper sanitization. Ingestion points: System state data retrieved via PowerShell in SKILL.md. Boundary markers: Safety warnings are present in the documentation but are not programmatically enforced in the provided scripts. Capability inventory: Includes process termination, service modification, and registry deletion. Sanitization: No explicit escaping or filtering is performed on retrieved data before it is used in subsequent operations.

pc-ops