Skip to content

AGENTS.md

Latest commit

 

History

History
49 lines (35 loc) · 2.22 KB

AGENTS.md

File metadata and controls

49 lines (35 loc) · 2.22 KB
Raw

AGENTS.md

Project: ManagedCode.MCPGateway.Benchmarks Owned by: ManagedCode.MCPGateway package maintainers

Parent: ../../AGENTS.md

Purpose

  • This project owns BenchmarkDotNet performance and allocation benchmarks for ManagedCode.MCPGateway.
  • It measures representative package hot paths without becoming a shipping runtime dependency.

Entry Points

  • Program.cs
  • BenchmarkCatalog.cs
  • McpGatewayIndexBuildBenchmarks.cs
  • McpGatewaySearchBenchmarks.cs
  • McpGatewayToolSetBenchmarks.cs

Boundaries

  • In scope: benchmark scenarios, benchmark catalog fixtures, allocation/runtime measurement, and benchmark documentation.
  • Out of scope: production runtime code, integration tests, CI release gates, and test-only assertion helpers.
  • Protected or high-risk areas: benchmark realism, stable input data, Release-only execution, and avoiding benchmark-only changes in production code.

Project Commands

  • benchmark: dotnet run -c Release --project benchmarks/ManagedCode.MCPGateway.Benchmarks/ManagedCode.MCPGateway.Benchmarks.csproj -- --filter "*"
  • benchmark-search: dotnet run -c Release --project benchmarks/ManagedCode.MCPGateway.Benchmarks/ManagedCode.MCPGateway.Benchmarks.csproj -- --filter "*Search*"
  • benchmark-index: dotnet run -c Release --project benchmarks/ManagedCode.MCPGateway.Benchmarks/ManagedCode.MCPGateway.Benchmarks.csproj -- --filter "*Index*"
  • benchmark-tools: dotnet run -c Release --project benchmarks/ManagedCode.MCPGateway.Benchmarks/ManagedCode.MCPGateway.Benchmarks.csproj -- --filter "*ToolSet*"

Applicable Skills

  • mcaf-dotnet
  • mcaf-dotnet-profiling
  • mcaf-dotnet-format
  • mcaf-dotnet-roslynator
  • mcaf-solid-maintainability

Local Constraints

  • Run benchmarks in Release.
  • CI and release benchmark coverage run the complete BenchmarkDotNet suite with --filter "*"; do not replace it with smoke-only, dry-run, or short-run benchmark coverage.
  • Keep benchmark fixtures deterministic and network-free.
  • Keep benchmark methods focused on one measurable operation.
  • Use BenchmarkDotNet MemoryDiagnoser so allocations are visible with mean timing.
  • Do not optimize production code only for a synthetic benchmark shape; the scenario must represent a real gateway path.