# PromptKit — Versioned Prompt Library for Engineers

> PromptKit is a versioned prompt library with 157 components and 3 interactive workflows to bootstrap, evolve, and maintain engineered prompts from the CLI.

## Install

Paste the prompt below into your AI tool:

## Quick Use

```bash
npx @alan-jowett/promptkit
npx @alan-jowett/promptkit list --all
npx @alan-jowett/promptkit search "memory safety"
npx @alan-jowett/promptkit show review-cpp-code
npx @alan-jowett/promptkit list
```

## Intro

PromptKit is a versioned prompt library with 157 components and 3 interactive workflows to bootstrap, evolve, and maintain engineered prompts from the CLI.

**Best for:** teams who want repeatable prompts for reviews, specs, and audits

**Works with:** Node.js 18+, GitHub Copilot CLI, Claude Code, OpenAI Codex CLI

**Setup time:** 2-6 minutes

### Key facts (verified)

- GitHub: 57 stars · 11 forks · pushed 2026-05-12.
- License: MIT · owner avatar + repo URL verified via GitHub API.
- README-verified entrypoint: `npx @alan-jowett/promptkit`.

## Main

- Use PromptKit’s “Bootstrap → Evolve → Maintain” lifecycle to keep prompts and requirements aligned as the codebase changes.

- Treat `list --all`, `search`, and `show` as your catalog workflow: discover a template, inspect cross-references, then run it in a fresh LLM session.

- When using Claude Code, follow the README’s manual bootstrap approach (`Read and execute bootstrap.md`) instead of expecting CLI skills to run.

### Source-backed notes

- README states the library contains 157 components (personas, protocols, formats, taxonomies, templates) across 4 pipelines.
- README’s Quick Start shows `npx @alan-jowett/promptkit` plus catalog commands like `list --all`, `search`, and `show`.
- README lists three interactive workflows with entry points under `templates/` (spec extraction, engineering, maintenance).

### FAQ

- **Do I need to clone the repo?**: No — README recommends `npx @alan-jowett/promptkit` for zero-clone usage; cloning is optional.
- **Which LLM tool does it work with?**: README lists GitHub Copilot CLI, Claude Code, and OpenAI Codex CLI; manual mode is also documented.
- **What’s the simplest first run?**: Run `npx @alan-jowett/promptkit`, then use `list --all` and `search` to find a template for your task.

## Source & Thanks

> Source: https://github.com/microsoft/PromptKit
> License: MIT
> GitHub stars: 57 · forks: 11

---

<!-- ZH -->

## Quick Use

```bash
npx @alan-jowett/promptkit
npx @alan-jowett/promptkit list --all
npx @alan-jowett/promptkit search "memory safety"
npx @alan-jowett/promptkit show review-cpp-code
npx @alan-jowett/promptkit list
```

## Intro

PromptKit 把 prompt 当工程资产：157 个组件 + 3 条交互工作流（Bootstrap/Evolve/Maintain），可用 CLI 组合、搜索、审计与复用模板，减少漂移。

**Best for:** 想把 prompt 变成可复用资产、做评审/规格/审计的团队

**Works with:** Node.js 18+；GitHub Copilot CLI / Claude Code / Codex CLI

**Setup time:** 2-6 minutes

### Key facts (verified)

- GitHub：57 stars · 11 forks；最近更新 2026-05-12。
- 许可证：MIT；作者头像与仓库链接均已通过 GitHub API 复核。
- README 中核对过的入口命令：`npx @alan-jowett/promptkit`。

## Main

- 用 PromptKit 的 “Bootstrap → Evolve → Maintain” 生命周期，让 prompt/规格/实现随着代码演进保持一致。

- 把 `list --all`、`search`、`show` 当成你的“组件目录”工作流：先检索模板，再查看引用关系，然后在新会话执行。

- 在 Claude Code 场景下按 README 的手动 bootstrap 方式（`Read and execute bootstrap.md`）使用，而不是指望 CLI skill 自动运行。

### Source-backed notes

- README 写明共有 157 个组件（personas/protocols/formats/taxonomies/templates），覆盖 4 条 pipeline。
- README 的 Quick Start 给出 `npx @alan-jowett/promptkit` 以及 `list --all`、`search`、`show` 等目录命令。
- README 列出三条交互工作流，并标注了 `templates/` 下的入口文件（spec extraction、engineering、maintenance）。

### FAQ

- **必须 clone 仓库吗？**：不必。README 推荐用 `npx @alan-jowett/promptkit` 零 clone 使用；clone 只是可选方案。
- **支持哪些 LLM 工具？**：README 列出了 GitHub Copilot CLI、Claude Code、OpenAI Codex CLI，并提供手动模式。
- **最简单的第一次使用？**：先运行 `npx @alan-jowett/promptkit`，再用 `list --all`/`search` 找到适合的模板。

## Source & Thanks

> Source: https://github.com/microsoft/PromptKit
> License: MIT
> GitHub stars: 57 · forks: 11


---
Source: https://tokrepo.com/en/workflows/promptkit-versioned-prompt-library-for-engineers
Author: Prompt Lab