supported-tools.md

English | 한국어 | 中文 | 日本語 | Español

Supported AI Tools

Codingbuddy works with multiple AI coding assistants through a unified rules system.

Overview

Tool	Integration	Setup Guide
Claude Code	MCP Server	Guide
Cursor	Rules Directory	Guide
GitHub Copilot / Codex	Instructions File	Guide
Antigravity	Config Directory	Guide
Amazon Q	Rules Directory	Guide
Kiro	Spec Directory	Guide
OpenCode	Rules Directory	Guide

Claude Code

Integration Type: MCP (Model Context Protocol) Server

Claude Code connects via MCP, providing full access to project configuration, rules, and specialist agents.

Quick Setup

1. Add to Claude Desktop config:

   macOS: ~/Library/Application Support/Claude/claude_desktop_config.json

   Windows: %APPDATA%\Claude\claude_desktop_config.json

   {
     "mcpServers": {
       "codingbuddy": {
         "command": "npx",
         "args": ["codingbuddy", "mcp"]
       }
     }
   }

2. Restart Claude Desktop

Features

• Full MCP resource access (config, rules, agents)
• Tool calls (search_rules, get_agent_details, parse_mode, recommend_skills)
• Prompt templates (activate_agent)

Full Guide

Cursor

Integration Type: Rules Directory

Cursor uses .cursor/rules/ for project-specific instructions.

Quick Setup

1. Create .cursor/rules/ directory
2. Reference common rules:

<!-- .cursor/rules/codingbuddy.md -->

# Project Rules

Follow the common rules in `packages/rules/.ai-rules/`:

- Workflow: @packages/rules/.ai-rules/rules/core.md
- Quality: @packages/rules/.ai-rules/rules/augmented-coding.md
- Context: @packages/rules/.ai-rules/rules/project.md

Features

• File reference with @ syntax
• Project-specific customizations
• Agent context via file references

Full Guide

GitHub Copilot / Codex

Integration Type: Instructions File

GitHub Copilot uses .github/copilot-instructions.md for custom instructions.

Quick Setup

1. Create instructions file:

<!-- .github/copilot-instructions.md -->

# Coding Standards

Follow the guidelines in `packages/rules/.ai-rules/rules/`:

## Workflow
Use PLAN → ACT → EVAL workflow as defined in core.md

## Code Quality
- TDD approach (Red → Green → Refactor)
- TypeScript strict mode
- 80%+ test coverage

Features

• Markdown-based instructions
• Repository-wide settings
• Team-shared configuration

Full Guide

Antigravity

Integration Type: Config Directory

Antigravity (Gemini-based) uses .antigravity/ for configuration.

Quick Setup

1. Create .antigravity/rules/ directory
2. Add rule references:

<!-- .antigravity/rules/project.md -->

# Project Guidelines

Reference: packages/rules/.ai-rules/rules/core.md
Reference: packages/rules/.ai-rules/rules/augmented-coding.md

Features

• Gemini model integration
• Rule file references
• Project context awareness

Full Guide

Amazon Q

Integration Type: Rules Directory

Amazon Q Developer uses .q/rules/ for custom rules.

Quick Setup

1. Create .q/rules/ directory
2. Add consolidated rules:

<!-- .q/rules/codingbuddy.md -->

# Development Standards

Follow packages/rules/.ai-rules/ for consistent coding practices.

Key files:
- packages/rules/.ai-rules/rules/core.md (workflow)
- packages/rules/.ai-rules/rules/augmented-coding.md (TDD)

Features

• AWS integration
• Enterprise features
• Custom rule support

Full Guide

Kiro

Integration Type: Spec Directory

Kiro uses .kiro/ for specifications and steering files.

Quick Setup

1. Create .kiro/steering/ directory
2. Add steering file:

<!-- .kiro/steering/codingbuddy.md -->

# Project Steering

Apply rules from packages/rules/.ai-rules/:
- Workflow modes (PLAN/ACT/EVAL)
- TDD development
- Code quality standards

Features

• Spec-driven development
• Steering file system
• Task management integration

Full Guide

OpenCode

Integration Type: JSON Configuration

OpenCode (and its successor Crush by Charm Bracelet) uses JSON configuration files with agent-based workflows.

Quick Setup

1. Create .opencode.json (or crush.json):

{
  "instructions": [
    "packages/rules/.ai-rules/rules/core.md",
    "packages/rules/.ai-rules/rules/augmented-coding.md"
  ],
  "agent": {
    "plan-mode": {
      "prompt": "{file:packages/rules/.ai-rules/agents/plan-mode.json}",
      "permission": { "edit": "deny" }
    },
    "act-mode": {
      "prompt": "{file:packages/rules/.ai-rules/agents/act-mode.json}",
      "permission": { "edit": "allow" }
    }
  },
  "mcp": {
    "codingbuddy": {
      "command": ["npx", "codingbuddy", "mcp"]
    }
  }
}

Features

• Terminal-native TUI interface
• Agent-based PLAN/ACT/EVAL workflow
• MCP server integration
• Fine-grained permission control

Full Guide

Adding New Tools

Codingbuddy is designed to support additional AI tools:

1. Create adapter guide in packages/rules/.ai-rules/adapters/{tool}.md
2. Create tool directory .{tool}/
3. Reference common rules from packages/rules/.ai-rules/

See Contributing for details.

Comparison

Feature	Claude	Cursor	Copilot	Antigravity	Q	Kiro	OpenCode
MCP Support	✅	❌	❌	❌	❌	❌	✅
File References	✅	✅	❌	✅	✅	✅	✅
Agent Activation	✅	⚠️	❌	⚠️	⚠️	⚠️	✅
Project Config	✅	⚠️	❌	⚠️	⚠️	⚠️	✅

✅ Full support | ⚠️ Partial (via file reference) | ❌ Not supported

Next Steps

• Getting Started - Initial setup
• Philosophy - Design principles
• API Reference - MCP capabilities