The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill defines an environment for processing and generating automation logic based on project files and UI interaction, creating a surface for indirect injection.
Ingestion points: Automation source files (.xaml, .cs), configuration files (project.json), and metadata files (EntitiesStore.json) entered via CLI or file system tools.
Boundary markers: Absent for raw file editing, but the skill enforces strict validation and build gates using the uip rpa CLI tools before tasks are declared finished.
Capability inventory: The skill documents activities for remote script execution on virtual machines (Azure RunScriptAsync, AWS RunPowerShellCommand), Excel macro/VBA invocation, and broad file system/network access via cloud provider integrations.
Sanitization: Relies on the RPA platform's internal validation and the agent's adherence to the phase-gated validation rule.
[COMMAND_EXECUTION]: The skill makes extensive use of the uip rpa CLI tool for project lifecycle management, including initialization, dependency resolution, validation, and execution of workflows.
[DYNAMIC_EXECUTION]: Provides instructions for activities that execute dynamic or external code, such as JavaScript within browser-based forms, Excel Macros, and VBA scripts. These are standard automation capabilities within the UiPath ecosystem.
[EXTERNAL_DOWNLOADS]: Documents and facilitates connections to established cloud services and providers (AWS, Azure, Google). These references point to official endpoints and are consistent with the skill's primary purpose of managing enterprise automation resources.

uipath-rpa