The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes dotnet commands to install templates and initialize new projects.
The scripts/interactive_init.py script utilizes subprocess.run with a list of arguments rather than a shell string, which effectively mitigates command injection risks.
Instructions in SKILL.md explicitly direct the agent to show a summary of parameters and obtain user confirmation before proceeding with command execution.
[EXTERNAL_DOWNLOADS]: The skill downloads and installs the NetCorePal.Template from the official NuGet registry.
The template is a legitimate development resource belonging to the skill's author, netcorepal.
The installation process uses standard .NET CLI tooling and official package registries, representing normal developer workflow.
[INDIRECT_PROMPT_INJECTION]: The skill accepts user-defined inputs for project naming and configuration, creating a surface for potential instruction injection through data.
Ingestion points: User-supplied values for parameters like ProjectName, OutputDir, and technical selections (Framework, Database) enter the skill's context via SKILL.md or CLI arguments to the script.
Boundary markers: Procedural boundaries are established by instructions requiring the agent to present a parameter summary for manual user review and confirmation.
Capability inventory: The skill possesses the capability to write to the local filesystem and install templates via subprocess calls to the dotnet utility.
Sanitization: The scripts/interactive_init.py script provides sanitization by normalizing the project name into PascalCase and validating technical choices (Framework, Database, MessageQueue) against hardcoded allow-lists, preventing arbitrary data from flowing into critical command segments.

cleanddd-dotnet-init