AI Briefing
KOSign in

AWS releases Aidlc: Adaptive workflow steering for AI coding agents across multiple harnesses

·2026.10.15 16:20

Key point

The open-source tool supports Claude Code, Cursor, GitHub Copilot, and other major AI coding environments with a unified, deterministic engine.

Details

Overview

AWS has released Aidlc (AI-Driven Development Life Cycle), an open-source tool designed to transform AI coding assistants into structured, verifiable software delivery workflows. The project, hosted at awslabs/aidlc-workflows under an MIT-0 license, provides a harness-neutral core that executes natively across various AI coding environments.

Supported Harnesses and Installation

Aidlc operates across a wide range of AI coding tools, including:

  • Claude Code
  • Kiro CLI and Kiro IDE
  • Codex CLI
  • Cursor
  • opencode
  • GitHub Copilot

Installation is streamlined for macOS, Linux, and WSL via a curl script that installs the native aidlc command and harness runtime without requiring Bun or Node.js. Windows users can install via PowerShell (irm ... | iex). The tool requires specific minimum versions for certain harnesses, such as Codex CLI >= 0.145.0 and GitHub Copilot CLI >= 1.0.74.

Core Architecture and Features

The system is built on a deterministic engine that ensures consistent behavior across all supported harnesses. Key architectural components include:

  • Workflow Structure: 5 phases and 33 stages covering the lifecycle from initialization to operation.
  • Agent System: 14 agents comprising 11 domain experts, 2 reviewers, and 1 adaptive composer.
  • Profiles: 11 workflow profiles tailored for features, bug fixes, infrastructure, security, POCs, and enterprise delivery.
  • Governance: Human approval gates, source-bound review evidence, and a 119-event audit trail for persistent state and team knowledge.

Development and Configuration

The repository structure separates the hand-authored, harness-neutral methodology in core/ from thin, harness-specific manifests in harness/<name>/. Developers can configure providers (including Amazon Bedrock) using aidlc config providers. The recommended model for optimal performance is Claude Opus 4.8. The project includes extensive testing suites and a aidlc doctor command for diagnosing configuration issues.

awsopen-sourceclaude-codegithub-copilotworkflow-automationdeveloper-toolsllm-agentsai-coding-tools
Read originalShare on Threads

This summary was generated automatically by AI. Check the original for the author's claims and context. Copyright belongs to the original author.

Our guide explains how the AI works. Report summary errors, attribution issues, or removal requests via Contact.

Related

Hacker NewsBerthd Releases Open-Source macOS Client for Persistent Remote Coding AgentsHacker NewsAgent Standup: Open-source tool visualizes Claude Code sub-agents as a live team meetingGeekNewsbuttonmasher: AI Agent Skill Simulating Impatient Users to Find Duplicate Action and Race Condition BugsProduct HuntFarol Launches: Open-Source macOS Terminal for Managing Parallel Claude Code and Codex AgentsGitHub BlogGitHub Copilot app for Beginners: How to teach Copilot about your codebaseAmazon ScienceAWS introduces priming to convert pre-trained Transformers into hybrid state space models with 99% less pretraining cost