The Agent Skills Directory

[DYNAMIC_EXECUTION]: The script scripts/Test-DotNetAssemblyLoad.ps1 generates a C# project and source file in a temporary directory and executes it via dotnet run. This is used to verify that Revit plugin assemblies can be loaded without metadata errors.
[DYNAMIC_EXECUTION]: The script scripts/Invoke-RevitWpfUiAutomation.ps1 uses Add-Type to compile C# code at runtime for Windows API interop (user32.dll), which is necessary for UI automation tasks like setting the cursor position and simulating mouse events.
[COMMAND_EXECUTION]: The skill utilizes several PowerShell scripts to automate Revit operations, including starting the Revit process (Start-RevitPluginSmoke.ps1), interacting with the UI (Invoke-RevitWpfUiAutomation.ps1), and performing build/pack operations with the .NET CLI (Test-RevitTemplatePackage.ps1).
[INDIRECT_PROMPT_INJECTION]: The skill defines an attack surface by ingesting Revit Journal files and plugin logs (e.g., in Get-RevitPluginEvidence.ps1). While this is required for verifying test results, it could process untrusted strings output by the plugin being tested.
Ingestion points: scripts/Get-RevitPluginEvidence.ps1 and scripts/Start-RevitPluginSmoke.ps1 read Revit Journal files (journal.*.txt) and plugin log files.
Boundary markers: The SKILL.md instructions emphasize identifying "expected evidence" and "fail on missing" switches, but do not provide explicit delimiters for the logs.
Capability inventory: The skill has the ability to execute shell commands, compile C# code, and simulate UI interactions.
Sanitization: No explicit sanitization or escaping of log content is performed; it is primarily used for pattern matching via Select-String.

revit-automate-test