Who is vibe-coding-playbook for?

Developers and team leads using Cursor or similar AI assistants, ideally on a TypeScript stack, who know npm and git basics.

What can you build with vibe-coding-playbook?

Add a Cursor rules file to a TypeScript project so the AI follows your conventions. Write feature requests as XML planning prompts and approve the plan before any code is generated. Run a verify script as a pre-commit hook to catch type, lint and test failures in AI-written code. Use the architecture and performance prompt templates to structure design discussions with an AI. Build a shared prompt library for your team based on the templates.

RiyaParikh0112 /

vibe-coding-playbook

Cursor rules, XML prompt templates and a verify script for AI-assisted coding

381 stars
3 stars since 3 Oct
Updated 13 months ago
Some experience
Learning & resource lists

Open on GitHub (opens in a new tab)

AI coding assistants often write code that ignores project patterns, skips error handling or adds technical debt. This playbook treats the AI like a junior engineer: it must propose a plan, wait for approval, and then have its output checked.

The kit is mostly documents. You copy a Cursor rules file and a folder of prompt templates into your project, then add a shell script that runs type checks, linting, circular-dependency detection and tests. Prompts use XML tags to separate the objective, context, constraints, deliverables and an approval gate.

Review has three layers: automated checks, AI-assisted checks of logic and edge cases, and a human look at architecture. The examples lean toward TypeScript, Next.js, tRPC and Prisma. The README's figures, such as 2-3x faster delivery and 50% fewer bugs, come with no supporting data, so treat them as claims.

What it does

Planning-first prompts
Each feature starts with an XML prompt that asks for a plan and tells the AI to wait for explicit approval before writing code.
Cursor rules file
A rules file sets protocols for TypeScript (no any), React conventions, API design with tRPC and Zod, security and performance.
Three-layer verification
Automated checks, AI-assisted functional review and human architectural review, with a sample bash script for the first layer.
Domain prompt templates
Templates for architectural decisions, legacy modernization, performance investigation, API design, databases and security.
Team adoption guidance
Suggestions for pilots, training, CI integration and shared prompt libraries across a team or organization.

What you could build

Add a Cursor rules file to a TypeScript project so the AI follows your conventions.
Write feature requests as XML planning prompts and approve the plan before any code is generated.
Run a verify script as a pre-commit hook to catch type, lint and test failures in AI-written code.
Use the architecture and performance prompt templates to structure design discussions with an AI.
Build a shared prompt library for your team based on the templates.

Get started

Clone the repo, copy the Cursor rules and the advanced-prompts folder into your project, then make the verify script executable and hook it into your npm scripts. Start your next feature with the planning-first XML prompt. Check the paths first, as the README's copy commands and script locations don't fully match.

git clone https://github.com/RiyaParikh0112/vibe-coding-playbook.git

cp .cursor/rules /your-project/.cursor/rules

cp -r advanced-prompts /your-project/.vibe/

Commands are copied from the project's README. Read it on GitHub for the full setup.