What are the different phases of ADW setup?

It offers three phases: Minimal (basic CLI prompts), Enhanced (SDK support and compound workflows), and Scaled (advanced state management, git operations, and worktree isolation).

Does it support issue trackers other than GitHub?

Yes, it includes abstractions for multiple trackers, including built-in support for GitHub and Beads (a local SQLite-based tracker), and can be extended for Jira or GitLab.

Does this work with existing projects?

Yes, the bootstrapper analyzes your project structure and intelligently adapts the ADW infrastructure to fit your specific conventions rather than using rigid templates.

What is an AI Developer Workflow (ADW)?

An ADW is a programmatic framework that allows AI agents to execute complex, multi-step tasks within your codebase autonomously, moving beyond simple chat interactions to structured execution.

Can I customize the workflows after bootstrapping?

Absolutely. The ADW architecture separates the agentic layer from your application code, encouraging you to modify templates and modules to suit your specific team requirements.

ADW Bootstrap

Name: ADW Bootstrap
Author: JoshuaOliphant

byJoshuaOliphant

•

생산성 및 워크플로

Bootstraps sophisticated AI agent orchestration infrastructure into any codebase for programmatic development and automated workflows.

This skill transforms any project into an AI-ready environment by setting up an AI Developer Workflow (ADW) layer. It enables programmatic agent orchestration via subprocesses or SDKs, allowing Claude to execute complex, multi-phase tasks like planning, TDD implementation, and automated deployment. With support for progressive enhancement—from basic prompt automation to scaled SDLC orchestration with state management and worktree isolation—it provides a structured framework for building with AI agents at scale while maintaining project-specific conventions.

주요 기능

01Multi-phase workflow automation including Plan, Implement, Test, and Deploy

022 GitHub stars

03Programmatic agent orchestration via subprocess and SDK execution

04Progressive setup phases: Minimal, Enhanced, and Scaled for evolving needs

05Structured observability with dedicated agent execution logs and state tracking

06Integrated support for issue trackers like GitHub and local SQLite-based systems

사용 사례

01Standardizing AI development patterns and agentic behaviors across a team's repositories

02Automating repetitive coding chores and feature implementations programmatically

03Scaling development by running multiple AI agents in parallel using worktree isolation

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add joshuaoliphant/claude-plugins adw-bootstrap

For use in Claude.ai and ChatGPT

Download Skill