Skip to main content

plan-a-pilot-program

Plan a pilot program with a bounded population, hypotheses, operational controls, success and harm measures, support, stop rules, and scale decision. Use when testing a new service, policy, product, process, or partnership before broader adoption.

来源信息

仓库
protosphinx/sphinxstack
最近来源活动
2026年7月29日 13:55
检测到的 SKILL.md 语言
英语
星标
3
Forks
0

安装方式

默认使用会先检查来源的 Prompt;你也可以切换为直接命令,或下载本地副本。

检查来源文件

决定是否安装前,请先阅读 SKILL.md,以及 SkillsMP 当前展示的配套文件。

文件资源管理器
2 个文件

正在显示 SKILL.md

SKILL.md
来源说明 · 只读预览
name
plan-a-pilot-program
category
start
description
Plan a pilot program with a bounded population, hypotheses, operational controls, success and harm measures, support, stop rules, and scale decision. Use when testing a new service, policy, product, process, or partnership before broader adoption.
# plan-a-pilot-program A pilot should reduce a named uncertainty without exposing participants to ungoverned risk. ## When to use - Use when representative real-world operation is needed before scale. - Do not select only easy participants or call a promotional launch an experiment. ## Procedure 1. State decision, hypotheses, unknowns, alternatives, sponsor, and authority. 2. Define representative population, inclusion, exclusion, recruitment, consent where needed, accessibility, and sample limitations. 3. Bound geography, time, volume, capability, staff, data, integrations, vendors, and financial exposure. 4. Establish baseline, success, quality, safety, fairness, privacy, cost, adoption, and harm measures. 5. Define workflow, training, support, incident, exception, continuity, rollback, and participant communication. 6. Set stop, pause, expand, repeat, modify, and reject criteria before starting. 7. Capture quantitative outcomes, qualitative evidence, operational burden, failures, and missing data. 8. Make the scale decision with limitations, conditions, owner, and next evidence. ## Done - A pilot plan records decision, population, scope, controls, measures, operations, support, stop rules, timeline, and owners - Representativeness, baseline, consent, accessibility, data, safety, support, measure, stop, and scale-decision checks verify learning
在 GitHub 查看