AI Collaboration Common Rules Released as Open Source

Gizin Inc. has published "AI Collaboration Common Rules v2.0" on GitHub Gist, systematizing practical know-how for AI-assisted development.

Background

Through development experience across multiple projects using Claude Code, we've identified common patterns for effective AI collaboration. By sharing this knowledge widely, we aim to contribute to the advancement of AI-assisted development.

Main Contents

1. Basic Principles

Information priority (System reality > Documentation)
Thorough self-investigation
Balanced proactivity
Proactive questioning principles

2. Proactive AI Collaboration

Confirmation for large-scope changes
Pre-consultation for data structure changes
Proposals for implementations differing from existing patterns
Decisions involving performance and security

3. Work Management

Efficient routine work with "mid-close" and "close"
Session management and documentation
Proper use of Todo and Task tools

4. Communication

Concise responses with the 4-line rule
Principles for avoiding verbose explanations

How to Use

Using Claude Code's @import feature, you can easily reference common rules across multiple projects:

markdown

# Project Name Operation & Improvement Guide
@../COMMON_RULES.md

## Project-Specific Rules
(Project-specific content follows)

Access

The common rules are published on the following GitHub Gist:

🔗 AI Collaboration Common Rules v2.0

Feedback and opinions are welcome via Gist comments or the developer's X (@HirokaKoizumi).

Future Development

These common rules are a "living document" that will be continuously improved through practice. Based on feedback from everyone, we plan to evolve them into more practical rules.

A detailed explanation article is available on AI TIPS:
Released Common Rules for AI-Assisted Development

We hope these common rules serve as valuable guidelines for everyone involved in AI-assisted development.