Executing Distributed System Tests

Runs the distributed systems test plans that the design skill produces. This is the execution half: it drives chaos scenarios, fault injection, partition tests, and consistency checks against your actual cluster or simulator. It won't improvise oracles or quietly skip steps that would let broken tests pass. Before reporting success it verifies the workload actually ran, the fault actually landed, and the checker actually observed something meaningful. Produces a structured findings report with a 10-state verdict system that distinguishes between real failures, flaky results, environmental problems, and gaps in test coverage. The process is deliberate: it catalogs your existing test tooling before writing anything new, asks about your environment capabilities up front, and runs boundary scenarios with separate verdicts per surface so an untested code path can't hide inside an aggregate pass.

Install to Claude Code

npx -y skills add shenli/distributed-system-testing --skill executing-distributed-system-tests --agent claude-code