The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes PowerShell and Python scripts that perform file system operations (read, glob) on user-provided directory paths (ExtensionPath, ConfigPath).
[OBFUSCATION]: The file scripts/cfe-diff.py contains multiple Unicode escape sequences (e.g., \u041f\u0435\u0440\u0435\u0434, \u0023\u0412\u0441\u0442\u0430\u0432\u043a\u0430). While these decode to standard 1C configuration keywords (Перед, #Вставка), the use of non-plain-text representations for logic-critical strings is a detected obfuscation pattern.
[INDIRECT_PROMPT_INJECTION]: The skill identifies a surface for indirect prompt injection.
Ingestion points: The scripts read XML configuration files and .bsl source code files from external paths provided via arguments.
Boundary markers: The output of the analysis (which includes code snippets from the processed files) does not appear to use boundary markers to distinguish external content from agent instructions.
Capability inventory: The skill uses powershell.exe and python3 to execute its logic, but the agent itself has access to Bash, Read, and Glob tools.
Sanitization: There is no evidence of sanitization or filtering of the content read from the external source files before it is presented to the agent.

cfe-diff