Skip to content

Latest commit

 

History

History
139 lines (96 loc) · 6.75 KB

File metadata and controls

139 lines (96 loc) · 6.75 KB

AI Coding Tools — The Practical Guide

Hands-on best practices for 9 AI coding tools — No hype, just what works. Prompt techniques, workflow design, multi-tool orchestration, and copy-paste configs to maximize your AI-assisted development.

简体中文 | English

GitHub stars License PRs Welcome

9 Tools
Full coverage
66 Tips
Claude Code deep-dive
7 Methodologies
Prompting / Debug / Test
Copy-paste Configs
Ready to use

Getting Started

New here? Three steps:

  1. Read Prompt Engineering — learn how to talk to AI coding tools
  2. Pick your tool → read its Quick Start section
  3. Read Task Decomposition — learn how to break work into AI-sized chunks

Already using AI tools? Jump to: Advanced Tips · Multi-Tool Workflows · Ecosystem


9 Tool Guides

Tool Type Highlights
Claude Code CLI Agent 66 tips, Agent + Skill + Hook workflows
Cursor IDE .cursorrules config, Composer Agent mode
GitHub Copilot IDE Plugin Inline completion + Agent mode + custom instructions
OpenClaw AI Agent Framework 338k Stars, multi-platform + Skills + Cron automation
Windsurf IDE Cascade Agent, automatic context
Gemini CLI CLI By Google, large codebase analysis
Kiro IDE By AWS, spec-driven development
Aider CLI Git-native, supports almost any LLM
Trae IDE By ByteDance, free Claude/GPT access

Every guide follows the same structure: Core Concepts → Quick Start → Prompt Tips → Advanced Usage → Config Templates


Universal Methodologies

These apply regardless of which tool you use:

Topic What it solves
Prompt Engineering Prompt techniques specific to AI coding — not generic prompt engineering
Task Decomposition Breaking large tasks into AI-manageable pieces
Code Review Best practices for AI-assisted code review
Debugging Systematic debugging methodology with AI
Context Management Managing the context window to prevent AI from "getting dumber"
Testing Strategy Writing tests with AI — approaches and pitfalls
Security Security risks in AI-assisted coding and how to mitigate them

Multi-Tool Workflows

The real productivity gains come from combining tools, not using one in isolation.

Workflow Description
Claude Code + Cursor Claude Code for architecture & complex refactoring, Cursor for daily coding
Claude Code + Copilot Claude Code for agentic tasks, Copilot for inline completions
Tool Selection Guide Which tool for which scenario — one table to rule them all

Ecosystem

These projects work together to cover the full AI-assisted development pipeline:

Learn the tools → Inject methodologies → Load expert roles → Orchestrate roles → Safety guardrails
(guide)           (superpowers)         (agents)            (orchestrator)      (shellward)
Project Purpose Description
This repo Learning Practical guides for 9 AI coding tools
superpowers-zh Methodology 20 skills for AI coding tools — TDD, debugging, code review, and more
agency-agents-zh Expert Roles 187 professional roles — turn AI into a security engineer, DBA, PM, etc.
agency-orchestrator Orchestration YAML-based multi-role collaboration for complex tasks
shellward Safety Prevent AI agents from running dangerous commands or leaking sensitive data

Full ecosystem setup guide →


Why This Guide?

Most AI coding resources fall into two camps: shallow "getting started" posts, or single-tool deep dives. This guide fills the gap:

  • Practical over theoretical — Every tip has been tested in real projects. No "AI will change everything" filler.
  • Cross-tool coverage — Compare approaches across 9 tools. Know when to use Cursor vs Claude Code vs Copilot.
  • Multi-tool workflows — The biggest wins come from combining tools. We show you how.
  • Copy-paste ready — Config templates, prompt snippets, and .cursorrules you can drop into your project today.
  • Prompt engineering for code — Generic prompt guides don't cover the nuances of AI coding. This one does.

Contributing

Contributions welcome! See CONTRIBUTING.md.

In short: add tips, fix outdated content, share workflow experiences — Issues and PRs both work.


Acknowledgments

This guide builds on excellent work from:


Make AI coding tools actually work for you.

Star this repo · Open an Issue · Contribute