The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill analyzes repository content (source code, configuration, and project files) to generate discovery reports and test infrastructure. This represents a surface for processing untrusted data; however, the skill mitigates this by requiring explicit user review and confirmation of the discovery report before generating any code.
[COMMAND_EXECUTION]: The skill utilizes standard .NET CLI commands (e.g., dotnet build, dotnet test, dotnet sln) and container runtime checks (docker info, podman info) which are necessary and expected for its primary purpose of setting up a development and testing environment.
[EXTERNAL_DOWNLOADS]: The instructions reference well-known and official sources, such as the NuGet API (nuget.org), for resolving package versions. These are recognized as trusted services for .NET development.
[DYNAMIC_EXECUTION]: The testing templates utilize WebApplicationFactory, a standard .NET library for hosting applications in-process during integration tests. This is a legitimate use case for dynamic execution within a controlled testing context.
[DATA_EXFILTRATION]: The skill includes proactive security measures, such as setting up a local HTTP proxy to a non-existent port (e.g., http://127.0.0.1:9) and using a BlockExternalHttpHandler to ensure that tests do not accidentally reach real external services.

dotnet-structured-testing