kotlinx-benchmark

Step 1: Define the measurement

1. DEFINE setup | benchmark implementation | target/profile change | comparison | regression investigation | failure diagnosis.
2. STATE one performance question, operation boundary, input distribution, mode, unit, baseline/candidate, and practical effect size before writing code.
3. IDENTIFY Kotlin/Gradle versions, project kind, benchmark source set/compilation, registered targets, runtime/plugin/JMH versions, host/runtime, report consumer, and existing benchmark conventions.
4. READ the current repository guide, latest release, and target-specific reference before version changes. Keep plugin and runtime versions identical.
5. TREAT the toolkit as Alpha and Wasm as experimental. A benchmark measures a controlled scenario; it does not explain production latency by itself.

Completion: hypothesis, metric, workload, target, toolchain, comparison method, and acceptance rule are explicit.

Step 2: Inspect the project

RUN from repository root:

python3 scripts/inspect-project.py --root . --json