reliability-reviewer

To analyze dependency failure modes, trace code paths at I/O boundaries to identify how external service failures manifest as hangs, crashes, or silent data loss. This ensures end-to-end traceability of failure propagation across your system.

Cascading failures and unbounded retries occur when code changes at I/O boundaries lack explicit named failure paths and proper recovery semantics. Identifying swallowed exceptions and unhandled errors helps prevent these reliability risks.

Review fault tolerance by applying a strict rule binding findings to explicit named failure paths, avoiding unbounded retries, and highlighting evidence gaps. This traces observable impacts when external services slow down or fail.

Perform a reliability review whenever a code change touches I/O boundaries, external services, or recovery semantics. This surfaces where dependency failures propagate and what observable impact results across the system.

Yes, the reliability reviewer identifies swallowed exceptions and unhandled errors by enforcing explicit named failure paths. It guides remediation by highlighting potential cascading risks and residual faults across dependencies.