vibe-coding-configuration

Bootstraps a consistent AI agent environment in any project. Run once per project to generate config files for Claude Code, Cursor, GitHub Copilot, and the OpenAI Agents spec (AGENTS.md).

• Creates .vibe/rules.md — project context (fill in after setup)
• Creates .vibe/mcp-triggers.md — tells agents when to use which MCP tools
• Creates .vibe/todo.md and .vibe/lessons.md — task tracking and self-correction log
• Generates CLAUDE.md, AGENTS.md, .github/copilot-instructions.md — agent instruction files
• Generates .cursor/mcp.json and .cursor/rules/ — Cursor IDE config
• Registers MCP servers with the Claude CLI

• Node.js (for npx MCP servers)
• Claude CLI installed and authenticated
• A .env file in the project root with:

  GITHUB_PERSONAL_ACCESS_TOKEN=your_token
  CONTEXT7_API_KEY=your_key

Linux / macOS

bash vibe-setup.sh

Windows (PowerShell, run from the project root)

.\vibe-setup-windows.ps1

> If PowerShell blocks execution, run: Set-ExecutionPolicy -Scope Process -ExecutionPolicy Bypass

The generated files contain TODO placeholders. Use one of the prompts below to fill them in. Copy the relevant prompt and paste it into your AI agent as your first message in the project.

Use this when running the bootstrapper on a codebase that already exists.

I've just run the vibe-setup script in this project. Several config files were generated with
TODO placeholders that need to be filled in based on the actual codebase. Your job is to analyze
the project and populate them. Do not guess — read the actual files.

Work through the following steps in order:

1. **Explore the codebase**
   - Read the directory structure (top-level and key subdirectories)
   - Read package.json, any lockfiles, config files (tsconfig, eslint, etc.)
   - Identify the stack: languages, frameworks, maj