# AI-DLC Workflows

> A harness-neutral, gated methodology and engine that turns AI coding assistants into structured, auditable software-delivery workflows.

AI-DLC (AI-Driven Development Life Cycle) Workflows is an open-source project from AWS Labs that runs a structured, gated software-development methodology inside AI coding harnesses. It installs an `aidlc` command and per-harness runtimes for Claude Code, Kiro CLI, Kiro IDE, Codex CLI, Cursor, opencode, and GitHub Copilot. The latest listed release is v2.11.1.

## What It Is

AI-DLC is a harness-neutral core that wraps AI coding assistants in a defined lifecycle so requirements, decisions, implementation, tests, and operational work stay connected. Users describe work with a command such as `/aidlc Build a REST API for inventory management`; the system selects a workflow from the request, asks for missing decisions, and stops at approval gates before moving forward. Codex CLI uses `$aidlc` instead.

## How the Workflow Is Structured

The README describes 5 phases and 33 stages from initialization through operation, with 14 agents (11 domain experts, 2 reviewers, and an adaptive composer) and 11 workflow profiles covering features, bug fixes, infrastructure, security, proofs of concept, and enterprise delivery. State is tracked persistently alongside team knowledge, learned rules, and a 121-event audit trail. The documentation also covers sensors, plugins, and a harness engineer guide for reshaping stages, agents, scopes, and rules.

## Setup Path

A native installer script for macOS, Linux, WSL, and Windows PowerShell installs the `aidlc` command and all harness runtimes without requiring Bun or Node.js. From the project root, `aidlc config --harness <name>` configures a harness and `aidlc doctor` checks the setup. A manual-copy path using Bun is available for those who cannot install a native executable. Model-provider setup belongs to the harness, and the README recommends capable reasoning models.

## Features
- Gated workflow with human approval gates
- 5 phases and 33 stages from initialization to operation
- 14 agents including domain experts, reviewers, and an adaptive composer
- 11 workflow profiles (Classic, Express, and focused workflows)
- 121-event audit trail with persistent state
- Team knowledge and learned rules
- Plugin mechanism for extra stages
- Native installer and `aidlc doctor` diagnostics
- Same deterministic engine across all supported harnesses

## Integrations
Claude Code, Kiro CLI, Kiro IDE, Codex CLI, Cursor, opencode, GitHub Copilot, Amazon Bedrock

## Platforms
WINDOWS, MACOS, LINUX, WEB, API, VSC_EXTENSION, CLI

## Pricing
Open Source

## Version
2.11.1

## Links
- Website: https://awslabs.github.io/aidlc-workflows/
- Documentation: https://awslabs.github.io/aidlc-workflows/guide/00-introduction/
- Repository: https://github.com/awslabs/aidlc-workflows
- EveryDev.ai: https://www.everydev.ai/tools/ai-dlc-workflows