The Agent Skills Directory

[SAFE]: The skill is designed for validating the structure of 1C configuration metadata. It performs various integrity checks on XML elements, GUIDs, and file paths.
[EXTERNAL_DOWNLOADS]: The Python implementation of the validation script (cf-validate.py) relies on the lxml library for efficient XML processing.
[COMMAND_EXECUTION]: The skill invokes local scripts (cf-validate.ps1 or cf-validate.py) to perform its primary function. This execution is scoped to the skill's purpose and uses standard arguments.
[PROMPT_INJECTION]: An indirect prompt injection surface is present as the skill parses external XML data provided by the user.
Ingestion points: The skill reads Configuration.xml, HomePageWorkArea.xml, and various language files within the specified configuration directory.
Boundary markers: No specific boundary markers or instructions to ignore embedded content are used during interpolation.
Capability inventory: The skill has the ability to write validation reports to the local filesystem if the OutFile parameter is provided.
Sanitization: The scripts perform structural validation (e.g., regex checks for GUIDs and identifiers), which provides a layer of sanitization for the data being processed.

cf-validate