ab-testing
Installation
SKILL.md
Ab Testing
Scope Boundaries
- Use when product, pricing, ranking, recommendation, or policy changes may impact user/business KPIs and causal validation is required.
- Use proactively when metric impact is uncertain, stakeholder opinions conflict, or ship/no-ship criteria are not explicit.
- Use when canary metrics alone are insufficient to support a decision.
- Do not use for deterministic functional verification; use
testing-*. - Do not use for long-term reliability telemetry design; use
observability-*.
Goal
Produce causally valid, operationally safe, and decision-ready experiment outcomes.