12 Experiment Design — Testing skill for Claude Code
Take a scored backlog idea and write the experiment doc before any assets — hypothesis, minimum viable test, tracking gate, pipeline row, and vendor briefs.
How to install 12 Experiment Design
Installs to ~/.claude/skills/chanl-ai-marketing-os-12-experiment-design/SKILL.md
mkdir -p ~/.claude/skills/chanl-ai-marketing-os-12-experiment-design && curl -fsSL https://raw.githubusercontent.com/chanl-ai/marketing-os/HEAD/skills/12-experiment-design.md -o ~/.claude/skills/chanl-ai-marketing-os-12-experiment-design/SKILL.md Restart Claude Code, or start a new session, for it to be picked up.
What 12 Experiment Design does
name: experiment-design description: Take a scored backlog idea and write the experiment doc before any assets — hypothesis, minimum viable test, tracking gate, pipeline row, and vendor briefs. Use when starting a growth test or when create skills refuse to run without an EXP-ID.
Experiment Design
Zoom in. Write the experiment *before* you build. The doc forces honesty on assumptions, metrics, and whether the test is even measurable.
When to use
When `growth-plan` has named some
Alternatives in Testing
- Deslop Reborn — One iteration of the IronClaw Reborn de-slop loop — take ONE Reborn crate, fan out parallel review sub-agents 12.6k ★
- GAAI Framework — 95+ Drop-in governance layer (.gaai/ folder) -- backlog authorization, cross-session memory, decision tracking 113 ★
- Solve — QA: solve levels 0 through N by creating minimum passing artifacts 92 ★
Full documentation available on GitHub
View Source RepositoryRelated Skills
Validate Hypothesis
Business hypothesis validation skill. Validates ideas through 6 phases (origin check, market confirmation, int
13 Experiment Review
Analyze a finished growth experiment — win or lose — update the pipeline, push next ideas to the backlog, syst
Ab Test Setup
When the user wants to plan, design, or implement an A/B test or experiment, or build a growth experimentation
LLMs Generator
Generate llms.txt (LLMs.txt v2 spec) and AI-discoverability packages for any project, so AI agents can find an
Cold Read Spec
Write the spec first, then prove a zero-context reader can build from it: numbered requirements with worked ex
Lead Research
[local test] Lead-research a company/domain/ICP → scored, cited lead profiles. Dev shim for the pro-research p
Related Agents
Biz MVP
Definition of the MVP (Minimum Viable Product). Use to identify essential features and plan the launch.
Repair Agent
Repairs a failed pipeline implementation. Use when a pipeline teammate reports PIPELINE_FAILURE and the implem
Engineering Minimal Change Engineer
Engineering specialist focused on minimum-viable diffs — fixes only what was asked, refuses scope creep, prefe