AI-Driven Life Cycle (AI-DLC) provides an adaptive workflow steering system for AI coding agents, addressing the need for structured, quality-assured, and human-controlled AI-assisted software development. It targets developers seeking to integrate AI agents into their lifecycle while maintaining oversight and reproducibility. The primary benefit is a more predictable and manageable AI development process.

How It Works

AI-DLC is fundamentally a methodology emphasizing a single source of truth, reproducibility, tool agnosticism, and a human-in-the-loop approach for critical decisions. It employs a three-phase adaptive workflow (Inception, Construction, and future Operations) that dynamically adjusts execution based on the value and complexity of specific requests, ensuring efficiency and relevance. Core rules are defined in Markdown files, allowing for broad compatibility.

Quick Start & Requirements

• Primary install / run command: Download the latest release zip from the Releases page or clone the repository.
• Non-default prerequisites and dependencies: Requires one of the following supported AI coding platforms/tools to be installed: Kiro, Kiro CLI, Amazon Q Developer IDE Plugin, Cursor IDE, Cline VS Code Extension, Claude Code CLI, or GitHub Copilot.
• Setup: Involves copying the aws-aidlc-rules/ and aws-aidlc-rule-details/ directories into platform-specific locations within your project (e.g., .amazonq/rules/, .cursor/rules/, .clinerules/, .copilot/instructions.md, CLAUDE.md, COPILOT.md, or AGENTS.md).
• Links: AI-DLC Methodology Blog (AWS Blog), AI-DLC Method Definition Paper, and documentation for supported platforms.

Highlighted Details

• Adaptive Intelligence: Executes only stages that add value to a specific request, avoiding unnecessary steps.
• Context-Aware: Analyzes existing codebase and project complexity requirements to tailor workflow execution.
• Risk-Based: Provides more comprehensive treatment for complex changes and remains efficient for simpler tasks.
• Question-Driven: Utilizes structured, multiple-choice questions within files for interaction, rather than relying solely on chat interfaces.
• Always in Control: Requires explicit user review and approval for generated execution plans and individual phases.

Maintenance & Community

The repository includes CONTRIBUTING.md and CODE_OF_CONDUCT.md files, indicating a framework for contributions and community interaction. No specific community platforms (like Discord/Slack) or notable contributors/sponsorships are detailed in the README.

Licensing & Compatibility

This library is licensed under the MIT-0 License. This permissive license generally allows for commercial use, modification, and distribution with attribution, making it broadly compatible with closed-source projects.

Limitations & Caveats

The "Operations" phase of the workflow is marked as "future" and not yet implemented. Setup requires integration with specific, pre-existing AI coding platforms, acting as a dependency. Troubleshooting indicates potential issues with rule loading, file encoding, session management, and platform-specific configurations, particularly on Windows where path handling may require forward slashes.