What problem does it solve? .NET applications suffer from latency regressions, memory growth, and throughput limits that are hard to fix without measurement. This Skill enforces a measure-first workflow so you fix the actual bottleneck — EF Core N+1 queries, sync-over-async, Gen2 GC pressure, thread-pool starvation — instead of guessing. ## Core Features & Use Cases - Measurement Tooling Guidance: Match symptoms to the right tool — BenchmarkDotNet for micro-benchmarks, dotnet-counters for live process stats, dotnet-trace and PerfView for CPU profiles, dotnet-gcdump for heap analysis. - Anti-Pattern Fixes: Concrete before/after C# code for EF Core N+1, sync-over-async, unbounded queries, HttpClient misuse, allocation hotspots, runtime-compiled Regex, and Kestrel misconfiguration. - Performance Budgets & CI Guards: Set enforceable budgets (p95 latency, allocated bytes per request, Gen2 rate) with a CI step that parses BenchmarkDotNet JSON and fails the build on regression. - Use Case: Your ASP.NET Core API's p95 latency jumped from 100 ms to 420 ms after a release. Use this Skill to profile with dotnet-trace, identify an EF Core N+1 introduced in the new endpoint, fix it with Include and AsSplitQuery, verify the delta with BenchmarkDotNet, and add a CI budget to prevent recurrence. ## Quick Start Ask the AI to profile your slow .NET endpoint and fix the identified bottleneck using the measure-identify-fix-verify-guard workflow.
