The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes external 1C:Enterprise binaries (1cv8.exe or ibcmd.exe) to perform its primary function of database creation. The execution logic is well-contained and incorporates security controls such as input sanitization and handle management to ensure safe process lifecycle.
[COMMAND_EXECUTION]: The scripts implement a redaction mechanism that identifies and masks password-related command-line flags (e.g., /P, --password) before printing command details to the console, preventing accidental credential exposure in logs.
[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from .v8-project.json files to configure its execution context.
Ingestion points: Project-level .v8-project.json files are read in scripts/db-create.ps1 and scripts/db-create.py.
Boundary markers: None identified.
Capability inventory: Command execution capabilities for interacting with the 1C platform.
Sanitization: The skill applies significant sanitization, including checking for quote characters in paths to prevent injection and validating user-supplied arguments against a blacklist of skill-controlled keys to prevent logic overrides.
[DYNAMIC_EXECUTION]: The PowerShell component utilizes Add-Type to compile an embedded C# helper class at runtime. This helper is designed to manage handle inheritance for spawned processes, which is a common and legitimate practice for robust process automation in Windows environments.

db-create