The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes dotnet build and dotnet test commands in its verification step (Step 6) to ensure the generated code is functional.
[REMOTE_CODE_EXECUTION]: The skill dynamically generates C# source code and integration tests which are then compiled and executed via dotnet test. This process involves running code derived from user inputs, which could be exploited to run unintended commands if the input is malicious.
[EXTERNAL_DOWNLOADS]: The integration test template in the architecture reference pulls the official postgres:17 image from a public container registry via Testcontainers. This is documented as a reference to a well-known service.
[PROMPT_INJECTION]: The skill serves as a surface for indirect prompt injection because user-provided feature names and schema details are interpolated into code templates without explicit sanitization. * Ingestion points: User-supplied names and entity fields provided during the scaffold command. * Boundary markers: The skill confirms scope with the user but lacks technical delimiters or warnings in the templates. * Capability inventory: File system writes (scaffolding), dotnet build, and dotnet test (SKILL.md). * Sanitization: No sanitization or validation of user-provided strings is mentioned.

scaffold