Experiment Designer with AI

# Experiment Designer — SoloStack Skill File
> Drop this file into your SoloStack repo and Claude Code will use it as
> context to run this workflow automatically.
---
## Overview
Design product experiments end-to-end — hypothesis to decision rule. Covers A/B tests, feature betas, pilots, fake-door tests, and pricing tests.
## Capabilities
1. **Hypothesis writing** — Forces the hypothesis into a falsifiable shape: "If X, then Y, measured by Z."
2. **Experiment-type selection** — Helps you pick A/B test, fake-door, pilot, beta, painted door, or qualitative — based on the question you're asking.
3. **Sample size guidance** — Estimates how many users you need to detect a meaningful effect. Saves you from running underpowered tests.
4. **Decision rule (pre-committed)** — You agree in advance what each possible result means — ship, kill, iterate, dig deeper.
5. **Risk + ethics flagging** — Surfaces ethical concerns (pricing tests with vulnerable users, dark patterns) and rollout risks before you commit.
## Pipeline
## Use Cases
- **A/B test a feature variant:** Design the test with proper sample size, decision rule, and guardrails before turning the flag on.
- **Pricing test:** Test a new pricing model with the right segment, control group, and exit criteria.
- **Fake-door MVP:** Validate demand for a feature before building it — with a clear "yes/no" threshold for "build the real thing."
- **Pilot or beta program:** Design the success criteria and timeline up front so the pilot can actually end.
## Warnings
## Sample Output
```json
{}
```
Download the complete skill file with full implementation details,
agent prompts, and ready-to-run scripts.