pm-abtest

pm-abtest is a skill for Claude Code, Codex from konglong87/superPM. It costs 59 tokens per session (2,722 once invoked), scanned A, original, MIT.

A guided process for planning an A/B test, which compares two versions of a product or page to measure whether a change causes a different result.

In plain words
What is it for?
Use it to define the test hypothesis, choose primary, supporting, and safety metrics, calculate sample size, and decide whether the experiment is ready.
Why use it?
It prevents teams from judging a change without a clear hypothesis, suitable measurements, or enough participants for a meaningful comparison.

Skill for Claude CodeCodex

Part of the super-pm plugin — 55 skills, 1 hook shipped together

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 skills/konglong87/superpm/pm-abtest
Any agent
npx skills add konglong87/superPM --skill pm-abtest
Clone the repo
git clone --depth 1 https://github.com/konglong87/superPM

Made for: Claude Code, Codex.

Or install super-pm, the plugin that ships this one along with the rest of its 55 skills, 1 hook.

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 pm-abtest

README.md
[![agentmods](https://agentmods.dev/badge/skills/konglong87/superpm/pm-abtest.svg)](https://agentmods.dev/skills/konglong87/superpm/pm-abtest)
Your own site
<a href="https://agentmods.dev/skills/konglong87/superpm/pm-abtest"><img src="https://agentmods.dev/badge/skills/konglong87/superpm/pm-abtest.svg" alt="Measured on agentmods" height="20"></a>
Per session 59 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 2,722 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. 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 $0.00059 $0.02722
Opus 5 $0.00030 $0.01361
Sonnet 5 $0.00012 $0.00544
Haiku 4.5 $0.00006 $0.00272

Measured 5d ago against content hash b12c7ab8bc03, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

pm-abtest 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 5d 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/03-growth-iteration/pm-abtest/SKILL.md · 332 lines

How it starts

The opening of the file, as written. The whole thing — 332 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Preamble

bash "$(dirname "${BASH_SOURCE[0]}")/../../check-update.sh" 2>/dev/null || true
mkdir -p docs/03-增长迭代/A-B测试

echo "🧪 A/B测试工具已启动"

# 检查数据指标体系
if [ -f "docs/02-方案设计/数据指标体系.md" ]; then
  echo "✅ 数据指标体系 - 已找到"
else
  echo "⏳ 数据指标体系 - 未找到"
fi

跨 Agent 交互规则

当流程要求与用户交互时:

  1. 如果当前环境支持 AskUserQuestion,使用 AskUserQuestion(最佳体验)。
  2. 如果当前环境不支持 AskUserQuestion,必须用普通聊天消息提出同样问题。
  3. 一次只问一个问题。
  4. 提问后必须停止当前回合,等待用户回答(STOP and WAIT)。
  5. 不得在用户回答前生成文档、写入 docs。
  6. 已有 docs 文件不能替代本轮用户回答。

执行流程

步骤 1: 定义测试假设

使用 AskUserQuestion 询问:

🎯 A/B测试假设设定

A/B测试需要明确的假设。请描述:

测试背景:为什么要进行这次测试? 示例:注册转化率低于行业平均水平(2% vs 行业5%)

测试假设:如果{改变什么},那么{预期结果},因为{原因}。 示例:如果将注册按钮从页面底部移到顶部,那么注册转化率将提升20%,因为用户更容易看到按钮。

请描述您的测试假设:

记录到变量 TEST_HYPOTHESIS


步骤 2: 设计实验方案

🔬 实验方案设计

实验变量

  • 对照组(Control):当前版本(现状)
  • 实验组(Treatment):{改动描述}

关键指标

  • 核心指标(Primary):{指标名称} - 直接影响业务结果
  • 辅助指标(Secondary):{指标名称} - 帮助理解变化原因
  • 护栏指标(Guardrail):{指标名称} - 确保不损害用户体验

示例(注册按钮测试)

  • 核心指标:注册转化率
  • 辅助指标:点击率、页面停留时长
  • 护栏指标:页面跳出率、用户满意度

请确认关键指标:

A) 指标合理,继续下一步 B) 需要调整核心指标 C) 需要补充辅助指标


步骤 3: 样本量计算

📊 样本量计算

需要的参数:

参数 说明 输入
基准转化率 对照组当前指标值 [X]%
最小可检测提升 期望的最小提升幅度 [X]%
显著性水平(α) 通常设为5%(0.05) 0.05
统计功效(1-β) 通常设为80%(0.8) 0.8

估算结果

  • 所需样本量(每组):约[X]个用户
  • 总样本量:约[X]个用户
  • 预估测试周期:约[X]天(基于当前日均流量)

样本量是否可行?

A) 可行,按此方案执行 B) 样本量过大,需要调整参数 C) 样本量太小,需要延长测试周期


步骤 4: 设定测试周期

⏱️ 测试周期设定

最小运行时间:{X}天(基于样本量计算) 建议运行时间:至少7天(覆盖工作日和周末) 最大运行时间:{X}天(避免环境变化影响)

运行规则

  • 流量分配:50%对照组 / 50%实验组
  • 用户分桶:按用户ID hash 随机分配
  • 互斥实验:确保同一用户不参与多个冲突实验

提前停止规则

  • 实验组指标显著优于对照组(p < 0.05)
  • 实验组护栏指标显著恶化
  • 出现严重技术问题

测试周期是否确认?


步骤 5: 数据收集与分析框架

📈 数据收集与分析:

数据收集

  • 埋点事件:{事件名称}
  • 数据存储:{存储位置}
  • 数据校验:每日检查数据完整性

分析框架

1. 数据清洗(去除异常值、测试账号)
2. 描述性统计(均值、标准差、分布)
3. 假设检验(t检验或Z检验)
4. 效应量计算(Cohen's d)
5. 分层分析(按用户特征分组)

决策规则

  • p < 0.05 → 统计显著,考虑上线
  • p ≥ 0.05 → 统计不显著,维持现状或重新设计
  • 护栏指标恶化 → 无论显著性如何,谨慎上线

Read the full file on GitHub · 332 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. 5d ago First seen · 332 lines · 59 tokens per session scan A b12c7ab8bc03

Subscribe to this mod's changes

pm-abtest is a skill published in the GitHub repository konglong87/superPM (61 stars, last pushed yesterday), licensed MIT. It adds 59 tokens to every session and 2,722 once invoked, about $0.0003 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 2026-08-30.

Related

Other skills, from other repositories

multi-search

智能多引擎搜索,自动检测网络环境并按优先级切换:DuckDuckGo -> Tavily -> Bing API -> Bing爬虫。支持自动配额管理和网络缓存。Invoke when user needs web search with automatic engine selection and network adaptation.

Nex-ZMH/Agent-websearch-skill · 60 tokens

freeride

Manages free AI models from OpenRouter for OpenClaw. Automatically ranks models by quality, configures fallbacks for rate-limit handling, and updates openclaw.json. Use when the user mentions free AI, OpenRouter, model switching, rate limits, or wants to reduce AI costs.

Shaivpidadi/FreeRide · 63 tokens

deepsafe-scan

Preflight security scanner for AI coding agents — scans deployment config, skills/MCP servers, memory/sessions, and AI agent config files (hooks injection) for secrets, PII, prompt injection, and dangerous patterns. Runs 4 model behavior probes (persuasion, sandbagging, deception, hallucination). Supports LLM-enhanced…

XiaoYiWeio/deepsafe-scan · 120 tokens

mapick

Mapick — Skill recommendation & privacy protection for OpenClaw. Scans your local skills, suggests what you're missing, and keeps other skills from seeing your sensitive data.

mapick-ai/mapick · 37 tokens

douyin-upload-mcp-skill

通过抖音创作者平台(creator.douyin.com)发布视频和图文内容。首选调用 MCP 工具完成操作,MCP 无法满足时可运行 Skill 脚本兜底。禁止自行启动外部浏览器访问抖音。.

WJZ-P/douyin-upload-mcp-skill · 67 tokens

team-tasks

Coordinate multi-agent development pipelines using shared JSON task files. Use when dispatching work across dev team agents (code-agent, test-agent, docs-agent, monitor-bot), tracking pipeline progress, or running sequential/parallel workflows. Covers project init, task assignment, status tracking, agent dispatch via…

win4r/team-tasks · 88 tokens