🚀 Anti-Chaotic Agent Kit

The Unified Agent Kit for AI-Driven Development - A comprehensive toolkit for software development powered by AI Agents.

🎯 Introduction

Anti-Chaotic is a comprehensive Agent Kit for Antigravity, designed to standardize and optimize your software development workflow.

🎯 Core Philosophy

Process over Speed: This kit is designed for development teams and collaborative environments. It prioritizes robust processes, structure, and long-term maintainability over "fast shipping" or quick hacks. It aims to bring standard engineering practices to AI-assisted coding.

🧩 Concepts

• Skills = Knowledge: What the agent knows (Best practices, languages, patterns).
• Rules & Workflows = Process: How the agent executes (Steps, restrictions, output formats).

We encourage teams to customize these skills and define their own rules & workflows to align with their specific enterprise standards.

📦 Components

• 🧠 12+ Multi-domain AI Skills - From Product Manager, Business Analyst to Lead Architect.
• 🔄 11 Automated Workflows - Pre-defined, reusable work processes.
• 📜 Rules Engine - A rule system that ensures AI Agents follow project standards.
• 📚 References Library - Documentation references for various technologies.

✨ Key Features

🎭 AI Skills (12 Virtual Experts)

| Skill | Description | | ----------------------- | ----------------------------------------------------- | | product-manager | Product management, roadmap, RICE/Kano prioritization | | business-analysis | Business analysis, user stories, use cases | | lead-architect | System architecture, ADR, RFC, technical specs | | designer | UI/UX design, design systems, accessibility | | frontend-developer | React, Vue, Angular, performance optimization | | backend-developer | Node, Python, Go, Java - API design, security | | devops-engineer | CI/CD, multi-cloud, observability | | qa-tester | Test planning, unit/E2E/security testing | | ai-engineer | RAG pipelines, prompt engineering, AI integration | | blockchain-engineer | Smart contracts, Web3, decentralized protocols | | rules-workflows | Create and manage rules, workflows | | skill-creator | Create new skills for the framework |

📜 Smart Rules Engine

• Documentation Rules: Standard document structure and formatting
• Clean Code Rules: Coding standards and best practices
• Project Rules: Project-specific conventions

ℹ️ Documentation & Tools: The default generated documentation is optimized for Obsidian. If your team uses Jira, GitHub, or other tools, simply connect the relevant MCP Server and customize the documents.md rule to align the output format.

🚀 Quick Start

1. Recommended Setup (MCP Servers)

To achieve the best results, we highly recommend installing the following MCP (Model Context Protocol) servers:

• sequence-thinking: AI model for advanced problem breakdown and step-by-step reasoning.
• context7: For managing and retrieving broad project context.

2. Initialize in your project

Run the following command in your project root to install the Agent Kit:

npx @kienha/anti-chaotic init

This will download the .agent folder containing all skills, rules, and workflows into your project.

3. Using the Agent Kit

Once initialized, start a new chat with your AI IDE (Cursor, Windsurf, etc.) and reference the installed rules/workflows.

Start a new project with AI:

/brainstorm Create a sales management application for a fashion store

Generate docs for an existing project:

/documentation

Design UI from requirements:

/ui-ux-design docs/020-Requirements/PRD-YourProject.md

4. Updating the Kit

To update your Agent Kit to the latest version:

npx @kienha/anti-chaotic update

📖 Workflow Guide

How to Use Workflows

Method 1: Slash Command

/brainstorm [Your project description]

Method 2: Direct Request

Please run the brainstorm workflow to analyze requirements for an e-commerce project

Method 3: File Reference

Read and execute the workflow at .agent/workflows/brainstorm.md

💡 Tips for Best Results

• Model Selection:
  • Workflows: Use Fast Mode with a reasoning model (Gemini 3 Pro High / Claude Thinking Model) when running flows.
  • Manual Edits: Use Planning Mode when making edits without a workflow.
  • Gemini 3 Flash: Use only for small tasks. Do not use for large, complex operations.
• Workflow Usage:
  • Tagging: Only tag a workflow in the first request of each session. Do not tag 2 or more workflows in the same session.
• Skills:
  • Explicit Activation: Proactively call a skill or use keywords associated with it for better efficiency.

🎓 Workflow Details

🔄 Automated Workflows

| Workflow | Description | Use Case | | :----------------------- | :--------------------------------------------------------------------------------------------------------- | :----------------------------------------------------------------- | | /bootstrap | Sets up project structure, installs dependencies, and configures environment based on architectural specs. | Start of Implementation Phase. | | /brainstorm | Analyze ideas with the user and create preliminary high-level documents (Roadmap, PRD). | Start of a new project or feature when you only have a rough idea. | | /break-tasks | Orchestrates breaking down requirements into actionable tasks for implementation. | When you have a PRD and need a task list. | | /custom-behavior | Safely customize Agent rules and workflows with impact analysis and user confirmation. | Adjust Agent behavior or fix recurring mistakes. | | /debug | Scientific debugging workflow: Hypothesize, Instrument, Reproduce, Analyze, Fix. | When facing complex bugs that need systematic analysis. | | /development | General coding workflow for implementing changes, bug fixes, or minor features. | Day-to-day coding tasks. | | /documentation | Generate comprehensive documentation (Architecture, API, Specs) from either Codebase or Requirements. | Onboarding or creating detailed specs. | | /gen-tests | Generate unit, E2E, security, and performance tests using the qa-tester skill. | Improving test coverage for new or existing code. | | /implement-feature | Orchestrates feature implementation from specification to deployment. | End-to-end feature development. | | /qa | Create comprehensive test case documents and test plans based on project requirements. | Planning testing strategy for a feature. | | /ui-ux-design | Transform requirements into comprehensive UI/UX design deliverables. | After requirements are finalized, before coding. |

🤝 Contributing

We welcome all contributions! See CONTRIBUTING.md for detailed guidelines.

How to Contribute

1. Fork the repository
2. Create a new branch (git checkout -b feature/amazing-feature)
3. Commit changes (git commit -m 'Add amazing feature')
4. Push to branch (git push origin feature/amazing-feature)
5. Open a Pull Request

📄 License

This project is released under the MIT License.