name	adws
description	Run autonomous development workflows using ADWS (Agentic Development Workflow System). Use for any multi-step development work including sprint implementation, bug fixes with verification, or phased feature development.
allowed-tools	Read, Write, Edit, Grep, Glob, Bash, Task, AskUserQuestion, TodoWrite

ADWS - Agentic Development Workflow System

Name: adws
Author: memyselfmike

This skill provides autonomous sprint execution using ADWS.

When to Use

Use this skill when:

Implementing a new epic or sprint
Multi-story development work
Bug fixes requiring verification
Any work requiring multiple phases
Phased feature development

Do NOT use when:

Single-line quick fixes (use direct edit)
Code exploration/research (use interactive Claude Code)
Simple documentation updates

Quick Start

# Initialize ADWS in project (if not done)
adws init

# Run workflow
adws run --input docs/vision.md

# Resume if interrupted
adws run --resume <workflow-id>

ADWS CLI Commands

Command	Description
`adws init`	Initialize ADWS in current project
`adws run --input <file>`	Start new workflow from input file
`adws run --resume <id>`	Resume interrupted workflow
`adws run --dry-run`	Preview workflow without executing
`adws status <id>`	Show workflow status
`adws list`	List all workflows
`adws phase <id> <name>`	Execute single phase
`adws clean --older-than 7d`	Clean old workflow state

Workflow Configuration

Create adws.config.yaml in your project:

phases:
  - name: plan
    commands:
      - command: /create-prd
        args: ["${input}"]
        model: sonnet
    outputs:
      - name: prd
        path: docs/prd.md
    gates:
      - name: prd_exists
        type: file_exists
        file_path: docs/prd.md

  - name: implement
    commands:
      - command: /implement-story
        model: sonnet
    gates:
      - name: tests_pass
        type: command
        command: pytest

Gate Types

Type	Description	Config
`file_exists`	Check file exists	`file_path: path/to/file`
`command`	Run shell command	`command: npm test`
`tests_pass`	Run test suite	`command: pytest`

Success Criteria

ADWS workflow is complete when:

All phases executed successfully
All hard gates passed
Workflow status shows "completed"
Output artifacts created in expected locations

adws

Install Skill

SKILL.md