The open-source home of AgentEval — the comprehensive .NET evaluation toolkit for AI agents.
AgentEval is the comprehensive .NET evaluation toolkit for AI agents — tool usage validation, RAG quality metrics, stochastic evaluation, behavioral policies, and model comparison — built first for Microsoft Agent Framework (MAF).
What RAGAS and DeepEval do for Python, AgentEval does for .NET.
| Repository | Description |
|---|---|
| AgentEval | The core evaluation toolkit — metrics, assertions, tracing, benchmarking |
| AgentEval.Cli | Command-line interface — evaluate from your terminal |
| AgentEval.Gatekeeper | AI-powered PR quality gate — every agent earns its merge |
dotnet add package AgentEval
// Evaluate whether your AI agent calls the right tools
result.ToolUsage!.Should()
.HaveCalledTool("SearchFlights")
.BeforeTool("BookFlight")
.And()
.HaveNoErrors();We welcome contributions! Check out each repository's CONTRIBUTING.md for guidelines.
Stop guessing if your AI agent works. Start proving it.
MIT Licensed · Built with ❤️ for the .NET community