AI-DLC Documentation
AI-DLC is a methodology — a structured, gated approach to AI-driven software
development (defined by AWS). This repository is its native, multi-harness
implementation: the methodology rendered as skills, agents, hooks, and tools
from one harness-neutral core/, so it runs natively in the CLI harness you use
— today Claude Code, Kiro CLI, Kiro IDE, Codex CLI, or opencode, and any capable CLI you port it to.
The methodology is the what; each harness distribution is the how for one
runtime, and every distribution is generated from the same source.
New here? The README has the install Quick Start and the "pick your harness" table. This page is the map of the documentation itself.
Three guides, one per reader
Pick by what you're trying to change:
| Guide | You are… | You change… |
|---|---|---|
| User Guide | building software with AI-DLC | nothing in the framework — you run /aidlc, answer at gates, review artifacts |
| Harness Engineer Guide | reshaping how AI-DLC behaves for your team | the data the framework reads: stages, agents, scopes, rules, sensors, knowledge — and porting to a new harness |
| Developer Reference | changing AI-DLC itself | the code that reads that data: the engine, hooks, CLI tools, the compile pipeline, the test suite |
The line between the Harness Engineer Guide and the Developer Reference is data versus code; the line between the User Guide and the rest is using versus shaping.
Running on a specific harness
The guides are harness-neutral; each harness's install steps and the handful of behaviours that differ live in Running on other harnesses (Claude Code is covered throughout the User Guide, whose examples run on it).
Building and contributing
Maintainers author in core/ and regenerate the dist/<harness>/ trees with
bun scripts/package.ts — see the Contributing Guide
for the full build-and-test loop, and Porting to a New Harness
to add one.