dotnet-slopwatch

To detect LLM reward hacking in .NET code changes, use Slopwatch to flag disabled tests, warning suppressions, empty catch blocks, arbitrary delays, and CPM bypass introduced during AI-driven edits.

Slop in C# source files and project files refers to shortcuts like disabled tests, empty catch blocks, and warning suppressions that degrade test quality, code safety, and correct behavior during LLM-driven development.

To distinguish new code issues from existing debt, initialize a baseline with Slopwatch before making changes, then run an analysis after LLM edits to identify only the newly introduced shortcuts and reward-hacking patterns.

Yes, Slopwatch works across API, Application, Domain, and Infrastructure layers, applying its analysis to C# source files, project files, and test suites to maintain code integrity during LLM-driven modifications.

To run a slopwatch analysis after an LLM generates code, execute the slopwatch analyze command to evaluate your .NET project against the established baseline for newly introduced shortcuts and reward-hacking behavior.

You need to install Slopwatch as either a local or global .NET tool to analyze .NET tests, which allows you to initialize a baseline and run subsequent analyses on your C# source files and project configurations.