// the take
Simmy bolts chaos-engineering primitives (fault, result-substitution, latency, custom behavior) onto Polly v7 policies so you can inject failure into real executions instead of mocking it. It's for teams already using Polly who want to verify their resilience policies actually do something under synthetic faults, in staging or even production.
The four policy types cover the practical cases (throw, return a bad result, add latency, run arbitrary code) and all come in sync/async and generic/non-generic flavors consistent with Polly's own API. Injection rate and enabled/disabled are both exposed as `Func<Context, ...>` delegates, so you can drive chaos from external config or target specific tagged operations without redeploying. Placing it innermost in a PolicyWrap is a genuinely good pattern - the surrounding retry/circuit-breaker/fallback policies you're actually testing see the fault exactly as they would a real one.
No commits since mid-2022, and Polly itself moved to the v8 ResiliencePipeline model in 2023 - Simmy was never ported and only works against the legacy IAsyncPolicy/Policy surface, so anyone on current Polly needs to pin an old version just to use it. There's no built-in blast-radius control beyond the injection-rate/enabled flag - no kill switch, no integration with feature-flag platforms, you build that yourself as shown in the separate demo repos. At 591 stars and 24 forks with a dead changelog, there's effectively nobody maintaining it if a future .NET or Polly change breaks it.
We read the README, the directory tree and a key source file in October 2026. Nobody pays to be listed. How we judge →