experiment-design-and-ab-testing

experiment-design-and-ab-testing is a skill for Claude Code from shennawardana23/skillme. It costs 157 tokens per session (2,302 once invoked), scanned A, original, Apache-2.0.

A guide to planning product A/B tests, where two versions are compared to learn which performs better. It covers hypotheses, test size, result interpretation, unexpected effects, and safety metrics.

In plain words
What is it for?
Use it to define a testable product question, choose the main metric and minimum worthwhile improvement, estimate the needed test size, and interpret results without repeatedly checking them or ignoring side effects.
Why use it?
It helps prevent teams from treating early or misleading results as proof that a change worked. It also separates deciding whether an outcome is real from checking whether the software works.

Skill for Claude Code

Written for Claude Code: shipped in a Claude Code plugin.

Part of the skillme plugin — 137 skills, 2 commands shipped together

Good fit Use it to define a testable product question, choose the main metric and minimum worthwhile improvement, estimate the needed test size, and interpret results without repeatedly checking them or ignoring side effects.

Compare 6 skills from other repositories ↓
Install with agentmods
npx agentmods add skills/shennawardana23/skillme/experiment-design-and-ab-testing
Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

Any agent
npx skills add shennawardana23/skillme --skill experiment-design-and-ab-testing
Clone the repo
git clone --depth 1 https://github.com/shennawardana23/skillme

Made for: Claude Code.

Or install skillme, the plugin that ships this one along with the rest of its 137 skills, 2 commands.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for experiment-design-and-ab-testing

README.md
[![agentmods](https://agentmods.dev/badge/skills/shennawardana23/skillme/experiment-design-and-ab-testing/github.svg)](https://agentmods.dev/skills/shennawardana23/skillme/experiment-design-and-ab-testing)
Your own site
<a href="https://agentmods.dev/skills/shennawardana23/skillme/experiment-design-and-ab-testing"><img src="https://agentmods.dev/badge/skills/shennawardana23/skillme/experiment-design-and-ab-testing/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

agentmods 80×15 button for experiment-design-and-ab-testing

Your own site · 80×15
<a href="https://agentmods.dev/skills/shennawardana23/skillme/experiment-design-and-ab-testing"><img src="https://agentmods.dev/badge/skills/shennawardana23/skillme/experiment-design-and-ab-testing.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 157 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 2,302 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5.1 $0.00157 $0.02302
Opus 5 $0.00078 $0.01151
Sonnet 5 $0.00031 $0.00460
Haiku 4.5 $0.00016 $0.00230

Measured 11d ago against content hash 5795582cdeab, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-11, from the pricing page.

Security

Grade A, and why

experiment-design-and-ab-testing scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 11d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

skills/experiment-design-and-ab-testing/SKILL.md · 182 lines

How it starts

The opening of the file, as written. The whole thing — 182 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Experiment Design and A/B Testing

This skill covers the product-decision layer of experimentation: what question the experiment answers, how big it needs to be, how to read the result honestly, and what else to watch besides the metric you're optimizing. It does not cover how to technically wire up a browser test, mock network calls, or automate a UI flow — for that, use skills/e2e-testing/ (Playwright page objects, CI config, flaky-test diagnosis). The two are complementary: e2e-testing verifies the code works before a rollout; this skill decides whether the rollout's outcome is real.

1. State a falsifiable hypothesis before running anything

Write the hypothesis in a form that could be proven wrong, before looking at any data:

"Changing X will change [primary metric] by at least [MDE] for [population], measured over [duration], because [causal mechanism]."

A hypothesis like "let's test a new checkout flow and see what happens" is not falsifiable — there's no result that could disprove it, so any outcome gets rationalized as a win after the fact. Committing to the metric, the minimum effect size, and the duration before the test starts is what prevents the analysis from being reverse-engineered to match whatever the data happened to show.

2. Minimum detectable effect (MDE) drives sample size — and the relationship is steep, not linear

MDE is the smallest true effect you want the test to be able to detect reliably. The qualitative shape of the tradeoff: required sample size grows roughly with the square of how small an effect you want to detect — halving the MDE you're targeting roughly quadruples the sample size needed, not doubles it. This is why "let's just detect any improvement, no matter how small" is not a free request — chasing a 0.5% lift on a metric that needs a 3% lift's worth of sample to detect reliably means the test will run far longer than the team expects, or will report "no significant difference" on an effect that's real but too small for the sample to see.

Read the full file on GitHub · 182 lines

Files

What ships with it

2 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 11d ago First seen · 182 lines · 157 tokens per session scan A 5795582cdeab

Subscribe to this mod's changes

experiment-design-and-ab-testing is a skill published in the GitHub repository shennawardana23/skillme (2 stars, last pushed 14d ago), licensed Apache-2.0. It adds 157 tokens to every session and 2,302 once invoked, about $0.0008 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.