I built ResilixForge, an open-source resilience toolkit for async Python services. It gives you the core failure-handling patterns as composable, declarative policies: - Retries with backoff - Timeouts - Circuit breakers - Bulkheads - Rate limits Instead of scattering try/except and retry logic across your codebase, you define policies once and compose them. Details: - Policy engine with no eval / no exec / no dynamic code execution - Full mypy --strict type checking - 200+ tests ...