App-vNext/Polly ? reverse-engineered prompt

Build me a .NET resilience library like Polly that helps apps handle flaky network and service calls more gracefully.

I want a clean fluent API where developers can create a pipeline and combine things like retry, timeout, circuit breaker, rate limiting, fallback, and hedging, then run their code through that pipeline. It should feel easy to use, thread safe, and work nicely with dependency injection too.

Please include a core package for the pipeline and strategies, extension support for DI and telemetry, and some testing helpers so the behavior is easy to verify. If you need to look up current .NET docs or best practices online, go ahead.

Make sure there are simple samples and clear README style usage examples so someone can copy the pattern into a real app right away.