# Benchmark Ab Runner

> Compare baseline and candidate query performance metrics and declare a benchmark winner.

- Skill: `rweisssieker-xp/benchmark-ab-runner` (Agent Skill)
- Install (CLI): `npx skillmds@latest add rweisssieker-xp/benchmark-ab-runner`
- Raw SKILL.md: https://api.skillmd.com/api/skills/rweisssieker-xp/benchmark-ab-runner/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Coding & Dev Tools
- Author: rweisssieker-xp (https://skillmd.com/u/rweisssieker-xp)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/rweisssieker-xp/benchmark-ab-runner

---


# Benchmark A/B Runner

Run:

```bash
node runtime/runTool.js benchmark_ab_runner '{"baseline":{"p95Ms":300,"cpuMs":80},"candidate":{"p95Ms":190,"cpuMs":55}}'
```

Returns deltas, winner, and verdict.