powershell-7-expert

Warn

Audited by Gen Agent Trust Hub on Sep 23, 2026

Risk Level: MEDIUMPERSISTENCEDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [PERSISTENCE]: The script scripts/crossplatform_automation.ps1 modifies shell profile files (~/.bashrc and ~/.zshrc) to persist environment variables when the Configure action is invoked. This is intended for platform configuration but uses persistent shell startup scripts.
  • [DYNAMIC_EXECUTION]: The TypeScript wrapper scripts/ps7_wrapper.ts dynamically builds and executes PowerShell commands using child_process.spawn. While it uses an argument array, the manual escaping of JSON strings for parameters adds complexity to command construction.
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests data from external API endpoints and user-supplied configuration data, providing a surface for indirect prompt injection if the external data contains malicious instructions.
  • Ingestion points: scripts/rest_api_consumer.ps1 (via Invoke-RestMethod) and scripts/crossplatform_automation.ps1 (via the $ConfigurationData parameter).
  • Boundary markers: No explicit boundary markers or instructions to ignore embedded commands were detected in the script logic.
  • Capability inventory: The skill possesses capabilities for file writing (Add-Content), network operations (Invoke-RestMethod), and container management (docker).
  • Sanitization: The TypeScript wrapper performs basic quote escaping for JSON objects passed as command-line parameters.
  • [COMMAND_EXECUTION]: The skill executes external commands including docker, pwsh, dpkg, and winget as part of its core automation and setup workflows.
  • [EXTERNAL_DOWNLOADS]: The documentation in references/ps7_quickstart.md provides instructions for downloading PowerShell installers directly from the official Microsoft PowerShell GitHub repository.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Sep 23, 2026, 08:06 AM
Security Audit — agent-trust-hub — powershell-7-expert