The Agent Skills Directory

[COMMAND_EXECUTION]: The skill relies on the dotnet CLI to compile, run, and manage packages for C# script files. It also utilizes system commands like chmod to modify file permissions for execution on Unix-like systems and mkdir for temporary project creation.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection (Category 8) because it ingests C# code or natural language intent to generate and execute scripts. If the agent is directed to process untrusted data (e.g., from a website or a PR) that contains malicious instructions, it may execute them as part of the script generation process.
Ingestion points: The skill takes 'C# code or intent' as a primary input, which is then written to a .cs file as shown in Step 2 of the workflow.
Boundary markers: The instructions do not specify any boundary markers or 'ignore embedded instructions' warnings for the input code.
Capability inventory: The skill uses dotnet for script execution and package management, chmod for changing file permissions, and mkdir/rm for filesystem operations.
Sanitization: There is no mention of sanitizing or validating the input code or intent before it is written to the filesystem and executed via the .NET SDK.

csharp-scripts