The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes a PowerShell script (meta-validate.ps1) to perform complex validation of 1C configuration objects. This is a legitimate use of command execution for a development tool.- [DATA_EXPOSURE]: The script reads XML data from the local filesystem based on user-provided paths and can write output to a user-specified file via the OutFile parameter. These operations are required for the skill's documented functionality of validating files and generating reports.- [INDIRECT_PROMPT_INJECTION]: The skill processes external XML data which represents a potential attack surface. Maliciously crafted XML could attempt to influence the agent's logic through the validation results.
Ingestion points: Reads XML files from the local filesystem using the ObjectPath parameter in scripts/meta-validate.ps1.
Boundary markers: The output uses clear headers (e.g., === Validation: ... ===) to separate the analysis from other context.
Capability inventory: The script has the capability to read any file accessible to the user and write to any path specified in OutFile.
Sanitization: The script performs strict pattern matching on identifiers and validates property values against allowed lists, though it does not explicitly disable XML external entity (XXE) resolution in the XmlDocument object.

meta-validate