design

A command for planning a Mixpanel experiment, which compares different versions of a product to measure their effects. It turns a hypothesis into metrics, traffic estimates, duration, and a draft configuration.

In plain words
What is it for?
For defining the hypothesis, success and safety metrics, required sample size, test duration, and statistical approach before an experiment begins.
Why use it?
It helps prevent launching an experiment with unclear goals, unsuitable measurements, or too little traffic. It stops at a draft so the launch remains a separate, deliberate action.

Command

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.

agentmods
npx agentmods add commands/mixpanel/ai-plugins/design
Clone the repo
git clone --depth 1 https://github.com/mixpanel/ai-plugins
Per session 0 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 3,104 The whole file, excluding the scripts and references it only reads on demand.
Security scan B 1 finding. Scan, not verified.
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 $0.00000 $0.03104
Opus 5 $0.00000 $0.01552
Sonnet 5 $0.00000 $0.00621
Haiku 4.5 $0.00000 $0.00310

Measured 2d ago against content hash a9110681b271, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade B, and why

design scanned grade B with 1 finding 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 2d 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.

Strips warnings and disclaimersmediumAnti-refusal

Omitting safety caveats hides risk from the user and is a common jailbreak preamble.

- Don't moralise about peeking — switch them to sequential.
plugins/mixpanel/skills/manage-experiment/commands/design.md · 172 lines

How it starts

The opening of the file, as written. The whole thing — 172 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Command: design

Design a Mixpanel experiment before launch. A well-designed experiment starts from the hypothesis and works backward: the hypothesis dictates the metrics that test it, the metrics dictate the sample size, the sample size + traffic dictate duration and testing model. This command stops at DRAFT — the irreversible launch happens in the separate launch command. Don't save the draft until the user explicitly confirms the configuration.

The umbrella SKILL.md defines the shared glossary (Variant, Primary/Guardrail/Secondary metric, Direction, Lift, MDE, CUPED, Winsorization, Multiple-testing correction). Phase-specific terms below.


Glossary (design-specific)

  • Hypothesis. A falsifiable, directional claim with a stated mechanism, bounded in time. Shape: "If <change>, then <metric> will <direction> by ≥<MDE>, because <mechanism>." Every other decision flows from this.
  • Power. The probability the experiment detects a true effect of size MDE. Default 80%.
  • Underpowered. Achievable MDE on available traffic exceeds the user's expected lift. Most likely outcome is "inconclusive"; reachable significance is biased upward (winner's curse).
  • Sequential vs Frequentist testing. Sequential makes peeking safe (boundary-based stopping); Frequentist requires a fixed sample committed up front. Most users should default to Sequential.

Components (design-specific)

Sizing formulas

Required sample per variant (two-sample, two-sided, 95% confidence, 80% power):

n = 16 × σ² / d²

Inverted for traffic-bound teams — the smallest effect detectable on available traffic (Kohavi's inversion):

MDE = 4σ / √n

The 16 is (z_{α/2} + z_β)² × 2 rounded. Variance σ² depends on metric type: Bernoulli p(1−p); Poisson ≈ mean; Gaussian computed from data. The full derivation, worked examples, lookup table, and the five remediations for underpowered experiments live in ../references/sizing.md.

Read the full file on GitHub · 172 lines

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. 2d ago First seen · 172 lines · 0 tokens per session scan B a9110681b271

Subscribe to this mod's changes

design is a command published in the GitHub repository mixpanel/ai-plugins (15 stars, last pushed 8d ago), licensed Apache-2.0. It costs nothing until one of its globs matches a file; then it loads 3,104 tokens. A static security scan graded it B with 1 finding (strips warnings and disclaimers). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.