awslabs/aidlc-workflows

AI‑DLC Workflows (awslabs/aidlc‑workflows)

What it is – AI‑DLC (AI‑Driven Development Life Cycle) is a set of rules that turn generative‑AI coding assistants (Kiro, Amazon Q, Cursor, Cline, Claude Code, GitHub Copilot, OpenAI Codex, etc.) into a structured, self‑correcting software‑development workflow. The repository ships a zip file (ai‑dlc‑rules‑vX.zip) containing two folders:

• aws-aidlc-rules/ – the core workflow definition (mostly markdown files that the supported IDE/agent reads as “rules” or “instructions”).
• aws-aidlc-rule-details/ – supplemental detail files referenced by the core rules.

When you copy these folders into the appropriate hidden directory for your chosen AI‑coding tool, the tool automatically applies the AI‑DLC workflow to every coding session. The workflow is described in a three‑phase adaptive process (inception → construction → operations) and is intended to keep code quality high, enforce standards, and let developers stay in control while the AI does the heavy lifting.

---

Quick‑start (common steps)

1. Download the latest release zip from the Releases page.
2. Extract it outside your project (e.g., ~/Downloads).
3. Copy the two sub‑folders into the hidden configuration directory required by your IDE/agent (see the platform‑specific sections below).
4. Verify that the rules appear in the IDE’s UI (screenshots are provided for each tool).
5. Code as usual – the AI assistant will now follow the AI‑DLC instructions, nudging you through the three‑phase workflow and offering self‑correction suggestions.

---

Supported platforms (setup snippets)

Platform	Where the rules live	Example command (macOS/Linux)
Kiro	.kiro/steering/ (core) and .kiro/ (details)	cp -R ~/Downloads/aidlc-rules/aws-aidlc-rules .kiro/steering/
Amazon Q	.amazonq/rules/ (core) and .amazonq/ (details)	cp -R ~/Downloads/aidlc-rules/aws-aidlc-rules .amazonq/rules/
Cursor	.cursor/rules/ai-dlc-workflow.mdc (or AGENTS.md)	cat > .cursor/rules/ai-dlc-workflow.mdc <<'EOF' … EOF
Cline	.clinerules/ (core) and .aidlc-rule-details/ (details)	cp ~/Downloads/aidlc-rules/aws-aidlc-rules/core-workflow.md .clinerules/
Claude Code	CLAUDE.md at project root (or .claude/CLAUDE.md)	cp ~/Downloads/aidlc-rules/aws-aidlc-rules/core-workflow.md ./CLAUDE.md
GitHub Copilot	.github/copilot-instructions.md	cp ~/Downloads/aidlc-rules/aws-aidlc-rules/core-workflow.md .github/copilot-instructions.md
OpenAI Codex	(similar to Copilot – uses AGENTS.md or custom instruction file)	see README

Each platform section also includes Windows PowerShell/CMD equivalents.

---

Core concepts

• Three‑Phase Adaptive Workflow – inception, construction, and operations. The rules guide the AI assistant through these stages, automatically adapting based on project context and developer feedback.
• Self‑correcting – The workflow includes verification steps that let the AI propose fixes and ask for confirmation before committing changes.
• Extensible – Additional rule sets can be dropped into the aws-aidlc-rule-details/ directory to customise behaviour for specific domains or compliance requirements.
• Tool‑agnostic – Because the rules are plain markdown (or simple config files), the same workflow can be used across many popular AI‑coding assistants.

---

Prerequisites

• A supported AI coding assistant installed and configured (Kiro, Amazon Q, Cursor, Cline, Claude Code, GitHub Copilot, or OpenAI Codex).
• Ability to run shell commands (or PowerShell/CMD on Windows) to copy the rule files into the hidden configuration directories.
• Optional: familiarity with the AI‑DLC methodology – see the linked blog post and the PDF specification for deeper background.

---

How to use

1. Install the assistant (e.g., add the Amazon Q VS Code extension).
2. Place the rule files as described for your tool.
3. Open your project in the IDE. The assistant will automatically load the core‑workflow.md (or equivalent) and start offering guidance.
4. Interact – the assistant may ask you to confirm a step, switch modes, or resolve a suggested correction. Follow the prompts; the workflow will keep track of progress.
5. Review – because generative AI can err, the README reminds you to review all AI‑generated code and monitor usage costs.

---

Extending the workflow

• Add new markdown files under aws-aidlc-rule-details/ (e.g., extensions/ for custom policies).
• Update the core core-workflow.md to reference the new files.
• The next time the IDE loads the rules, the extensions become active.

---

Supporting tools & resources

• Specification PDF – detailed formal description of the AI‑DLC 2.0 workflow.
• Blog post – high‑level overview of the methodology.
• Method Definition Paper – academic‑style write‑up of the underlying concepts.
• Troubleshooting – section in the README with common setup pitfalls.
• Version‑control recommendations – advice on how to integrate AI‑DLC with Git workflows.

---

License & contribution

• The project is released under the Apache‑2.0 license (see the LICENSE file).
• Contributions are welcomed via pull requests; see the Contributing section for guidelines.

---

TL;DR

AI‑DLC Workflows provide a ready‑made set of rule files that turn any major generative‑AI coding assistant into a disciplined, verifiable software‑development process. Install the assistant you prefer, copy the rule files into the tool’s hidden config folder, and let the AI guide you through an adaptive three‑phase workflow while you retain final control over every change.