Malcolm Jones (bossjones/Tony Dark) bossjones

BMad's Best Practices Cursor Custom Agents and Rules Generator

BMad Version: 3.1 (March 30, 2025) Cursor Version: 0.48+

Note from BMad: This repo has become much more successful than I originally envisioned, and I want to thank everyone for the feedback, support, contributions, and also feedback on the youtube channel. I now realize that many are using what I originally thought would be just a way to document some best practices around cursor rules and a simple rule that helps generate the rules - all with the intent of creating a separate updated workflow repo and set of documentation. But this has evolved.

With that in mind, going forward after this correction update - any major changes will follow a versioning standard, and upgrade paths or instructions to migrate if you are already using a prior version of rules or workflows whenever there is a significant potentially breaking change - although I will try to avoid that somehow depending on what comes with Cursor updates or workflow improvements. Keep

langchain LLMs-txt Overview - langchain-ai/mcpdoc - Expose llms-txt to IDEs for development vibe-code-benchmark - Code agents like Cursor have transformed how many of us work. Protocols like MCP (Model Context Protocol) can connect these agents with external data sources. This repo tests how different code agents compare and how best to connect them to external data.

  151  brew install go-task
  152  lvim Taskfile.yml
  153  task -g --list
  154  nvim --version
  155  asdf current
  157  cat ~/.config/lvim/config.lua | pbcopy
  158  cat ~/.config/lvim/config.lua
  160  ls ~/.local/share/nvim/site/pack/lazy/opt/nvim-treesitter
  161  ls ~/.local/share/lunarvim/site/pack/lazy/opt/

STRICT RULES:

🧠 DON'T BE LAZY AND BE ATTENTIVE! AND DON'T GET HALLUCINATIONS, BE CONSISTENT!
👨‍💻 Treat the user as a beginner web developer and you are super ultra expert professional AI assistant that will do all of the recommendations, suggestions, to control the workflow.
✅ Follow the user's requirements carefully & to the letter.
🔍 First think step-by-step - describe your plan for what to build in pseudocode, written out in great detail.
🔎 Search codebase first, then write code
🛠️ Fully implement all requested functionality.
🚫 Leave NO todo's, placeholders or missing pieces.
✔️ Ensure code is complete! Verify thoroughly finalized.

I'll analyze each of the six Python rules we created and provide a checklist of what's covered and what might be missing. Let me go through them one by one.

python-tdd-basics.mdc.md

✅ Covered:

Core TDD cycle explanation
Test structure template with type hints
Testing directory setup instructions
Basic test fixtures setup
Common testing patterns (fixtures, parametrized tests, exceptions)

params:

{ "repo_description": "A FastAPI application managed as a uv project, based on the multi-file example from the FastAPI documentation. It demonstrates a structured FastAPI application with routers, dependencies, and internal components.", "main_languages": "Python", "file_patterns": ".py, pyproject.toml, Dockerfile", "key_features": "FastAPI, API endpoints, dependency injection, router organization, uv package management" }

using your grep tools locate def prep_workspace (it's located in src/codegen_lab/prompt_library.py), I want you to enhance this to do the following:

You are an expert Python developer tasked with enhancing the prep_workspace function in our MCP-based tool. The current function prepares a workspace for cursor rules by creating directories and providing instructions. Your task is to expand this function to:

Write four mandatory cursor rule files to the client repo's cursor rules directory:
- tree.mdc.md
- repo_analyzer.mdc.md
- notify.mdc.md
- repomix.mdc.md

	#!/usr/bin/env python3

	"""
	Sequential Thinking FastMCP Server

	This module implements a FastMCP server for the Sequential Thinking tool, which enables
	dynamic and reflective problem-solving through a structured thinking process.

	The Sequential Thinking tool helps break down complex problems into manageable steps,
	allowing for revision, branching, and iterative refinement of thoughts. It maintains

	#!/usr/bin/env python3

	import asyncio
	import json
	import sys
	from typing import Dict, List, Any, Optional, TypedDict

	from mcp.server import Server
	from mcp.server.stdio import StdioServerTransport
	from mcp.types import CallToolRequestSchema, ListToolsRequestSchema, Tool

	"""Cursor Rules MCP Server.

	This module contains a Model Context Protocol (MCP) server that helps users create
	custom cursor rules for their repositories. The server analyzes repository structure
	and provides guidance for creating and organizing cursor rules.

	<thinking>
	# Implementation Checklist

	## Phase 1: Basic MCP Server Setup