The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill's reference materials and example scripts recommend the use of numerous well-known third-party libraries for automation, such as psutil, requests, paramiko, fabric, and pandas for Python, and chalk, ora, execa, and zx for Node.js. These are standard, reputable packages in their respective development ecosystems.
[COMMAND_EXECUTION]: The skill provides numerous utility scripts for system administration tasks like reporting, log analysis, and infrastructure provisioning. These scripts utilize standard CLI tools (aws, az, kubectl, curl) as intended for their administrative and educational purpose.
[DYNAMIC_EXECUTION]: The powershell/scripts/03-file-processor.ps1 script demonstrates the use of [scriptblock]::Create() to evaluate filter expressions. While this is a common technique for providing flexible query functionality in PowerShell, it represents a pattern where user-provided strings are executed as code within the environment.
[INDIRECT_PROMPT_INJECTION]: Several scripts in the collection are designed to ingest and process external data files (logs, CSV, JSON), which is a common pattern that nonetheless creates a potential attack surface.
Ingestion points: File reading operations in 03-csv-processor.py, 02-log-analyzer.sh, and 03-file-processor.ps1.
Boundary markers: The example scripts do not implement explicit delimiters or warnings to ignore instructions that might be embedded within the data files.
Capability inventory: The skill includes capabilities for file system modification, network requests (in API client examples), and subprocess execution.
Sanitization: The provided examples focus on functional parsing and do not include robust sanitization for adversarial content within the data files.

cli-scripting