A/B Testing
Design experiments with honest statistics.
Tested · Works
Test report
- Verdict
- Tested · Works
- Score
- Tested
- Jun 15, 2026
- Environment
- Claude Code 2.x
Designed an experiment with correct sample-size math and honest run-time estimates; flagged our proposed test as underpowered, which was right.
Scored on four weighted criteria — install, triggering, output vs. baseline, docs. How scoring works
- Installs cleanly 5/5
- Triggers reliably 4/5
- Output vs. baseline 8/10
- Docs & honesty 4/5
What it does
Experiment design, hypothesis framing, sample-size and duration math. Notably refuses to bless underpowered tests — the most common real-world A/B testing mistake.
Install
Copy the skill folder into ~/.claude/skills/ and describe the change you want to test.
Skills live in ~/.claude/skills/ (global) or .claude/skills/
(per-project). Restart Claude Code after installing.
Appears in