←Leaderboard/Agent Orchestration/moai-adk
Last commit on August 25, 2026·Created on September 16, 2025

modu-ai/moai-adk

“A spec-driven agentic development harness for structured coding workflows”
Combined rank
#282
across all skills
In Agent Orchestration
#16
category rank
Stars
1.2k
+0.6% in last 7d
Forks
226
0.0% in last 7d
Watchers
10
0.0% in last 7d
Traction scoreGitHub stars can be faked, so popularity alone can be misleading. Traction Score looks for broader signs of recent attention, adoption, and active maintenance.
TL;DR

The moai-adk provides a framework for implementing plan, run, and sync cycles within agentic coding environments. It is designed for builders creating automation tools that require strict adherence to specifications and high-performance execution.

By integrating domain patterns and Go best practices, the toolkit helps developers maintain consistency across complex codebase modifications while leveraging a zero-dependency binary architecture.

WHO IT'S FOR
AI agent / automation builders
implementing spec-driven agentic coding workflows
Backend / infrastructure engineers
developing high-performance Go CLI tools
Technical leads / architects
enforcing code quality gates for AI generation
OSS maintainers
managing multilingual documentation and README sync
Repository contents

93 skill files

Compatible AgentsThe repository documents support for these agents. The skills may also work with other agents that can load SKILL.md files, but they may need some setup or small changes.

Use this skill collection

Follow the documented setup, then try a first task.

Install and initialize on macOS, Linux, or WSL

Installs: MoAI-ADK CLI plus the generated Claude Code integration files for one initialized project · Claude Code

Before you start
  • Git, required on all platforms
  • Claude Code installed for the generated project integration

Platform: macOS, Linux, or WSL

Run from: Parent directory where the new project directory should be created

  1. Run this terminal command to install MoAI-ADK.

    curl -fsSL https://adk.mo.ai.kr/install.sh | bash
  2. In a terminal, initialize a project and enter its directory. Replace both occurrences of `my-project` with the desired project directory name.

    moai init my-project
    cd my-project
  3. After initialization, run Claude Code from the new project directory.

    claude        # or moai cc — run Claude Code inside the project
README.md ↗ · Checked Oct 2, 2026

Give it something to do.

Suggested first task

Review a small API surface

Uses moai-ref-api-patterns

  • Path to the API route, handler, or schema files to review
  • An initialized MoAI project running in Claude Code

Check the repository for this skill’s setup.

In this MoAI-initialized Claude Code project, delegate a read-only review of <path-to-api-files> through Agent(general-purpose) with backend instructions and the internal moai-ref-api-patterns reference. Check resource naming, pagination, request and response conventions, errors, and validation. Return evidence-backed issues and recommendations without editing files.
Suggested first task

Choose a React component pattern

Uses moai-ref-react-patterns

  • Path to one React or Next.js component
  • An initialized MoAI project running in Claude Code

Check the repository for this skill’s setup.

In this MoAI-initialized Claude Code project, delegate inspection of <path-to-component> through Agent(general-purpose) with frontend instructions and the internal moai-ref-react-patterns reference. Recommend whether compound components or custom-hook extraction fits, explain the state implications, and sketch a refactoring plan without editing files.
Suggested first task

Run a focused OWASP API review

Uses moai-ref-owasp-checklist

  • Path to the backend route, handler, middleware, or authorization files
  • An initialized MoAI project running in Claude Code

Check the repository for this skill’s setup.

/moai review --security Review <path-to-backend-files> using the internal moai-ref-owasp-checklist reference. Check object and function authorization, authentication, exposed properties, and resource consumption. Report evidence-backed findings and defenses without editing files.