This project introduces the GitHub Copilot Orchestra pattern, a multi-agent orchestration system designed to streamline AI-assisted software development. It addresses the challenge of maintaining code quality and test coverage while leveraging AI for rapid feature development. By enforcing Test-Driven Development (TDD) and implementing quality gates, it aims to combine AI coding speed with robust software engineering practices, benefiting developers seeking structured, test-driven AI workflows.

How It Works

The system employs a Conductor agent to orchestrate specialized subagents: Planning, Implementation, and Code Review. This multi-agent workflow enforces a strict Planning → Implementation → Review → Commit cycle for each development phase. The approach is advantageous as it breaks down complex tasks into manageable, iterative cycles, with each subagent having a dedicated context window to reduce AI hallucinations. TDD is strictly enforced during implementation, ensuring tests are written and pass before code is committed.

Quick Start & Requirements

• Primary Install/Run: Clone the repository and configure custom agents within VS Code Insiders.
• Prerequisites: VS Code Insiders (required for custom chat modes), an active GitHub Copilot Subscription, the GitHub Copilot Chat extension, and Git installed with the workspace initialized as a Git repository.
• Setup: Copy .agent.md files to VS Code's User Data/prompts directory or configure manually. Create a plans/ directory.
• Docs: VS Code Insiders download: https://code.visualstudio.com/insiders/

Highlighted Details

• Multi-Agent Workflow: Conductor orchestrates specialized Planning, Implementation, and Code Review subagents.
• TDD Enforcement: Strict adherence to writing failing tests first, then minimal code to pass.
• Quality Gates: Automated code review after each phase ensures standards are met before proceeding.
• Documentation Trail: Generates plan files and phase completion records for an audit trail.
• Mandatory Pause Points: User approval required for plan and phase commits, maintaining control.

Licensing & Compatibility

• License: MIT License.
• Compatibility: The permissive MIT license allows for commercial use and integration into closed-source projects without significant restrictions.

Limitations & Caveats

The system's heavy reliance on VS Code Insiders and the GitHub Copilot Chat extension presents a significant adoption barrier for users outside this specific development environment. The setup process for custom agents may also require a learning curve. While customizable, the default configuration depends on specific Claude AI models, necessitating access to those models.