The Agent Skills Directory

Input-Driven Command Construction: The skill incorporates user-provided test names and issue URLs into shell commands such as grep and dotnet run. This is a potential indirect prompt injection surface.
Ingestion points: User requests parsed in SKILL.md to extract test names and URLs.
Boundary markers: The skill lacks explicit delimiter-based boundary markers for user input when it is interpolated into shell commands.
Capability inventory: The skill uses grep, dotnet build, dotnet test, and dotnet run (which executes the QuarantineTools utility).
Sanitization: The skill relies on the agent's instructions to parse input into a specific structured format (Namespace.Type.Method), which acts as an implicit validation step.
Automated Code Refactoring: The skill automates the addition of attributes to source code using a specialized utility. The inclusion of mandatory build and test verification steps provides a safety layer to ensure that code changes are valid and do not introduce compilation errors.
Execution of Repository Utilities: The workflow utilizes local utilities found within the repository's tools directory. Running project-specific tools for maintenance tasks is an effective practice that ensures the automation aligns with the project's specific coding and safety standards.

test-management