start-working

Execute backlog tasks through a .task-board workflow with subagents and Playwright verification.

5|1|Updated Nov 20, 2025
One-click install
npx skills add https://github.com/andersnygaard/ettsted2 --skill start-working
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: start-working
Source: https://github.com/andersnygaard/ettsted2/tree/main/.claude/skills/start-working
Command: npx skills add https://github.com/andersnygaard/ettsted2 --skill start-working

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Managing the development workflow from task selection to implementation and verification, while adhering to project standards, can be complex and prone to manual errors. This Skill automates the entire process.

Core Features & Use Cases

  • Structured Workflow Enforcement: Guides task execution through backlogin-progressdone, ensuring adherence to project rules and best practices.
  • Automated Verification: Mandates Playwright CLI verification for frontend tasks, ensuring visual and functional correctness before marking work complete.
  • Use Case: When you're ready to implement the next feature, use this Skill to automatically pick the highest priority task, move it to "in-progress," guide the implementation, and ensure all verification steps are met.

Quick Start

Start working on the next priority task from the planning board.

Frequently Asked Questions about start-working

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I automate task execution through a development workflow?

Task automation executes backlog items through a structured workflow—backlog → in-progress → done—using subagents to implement and verify each task in sequence until completion. This Skill handles task selection, status updates, and verification enforcement automatically.

Can I verify frontend tasks automatically with Playwright during workflow execution?

Yes. Automated verification uses Playwright CLI to validate frontend tasks for visual and functional correctness before marking work complete, ensuring quality gates are met without manual intervention.

What's the best way to enforce project rules during autonomous task execution?

Project rules enforcement applies six workflow rules and project standards via CLAUDE.md context passed to subagents, prohibiting git commands and preferring built-in tools over Bash, maintaining consistency across all automated task execution.

Does this workflow automation work with task-board management systems?

This Skill applies specifically to projects using a .task-board workflow structure with backlog, in-progress, and done columns, processing tasks in priority order through autonomous subagent execution.

How do I handle complex task logic during automated execution?

Complex logic tasks route to sonnet subagents while simpler tasks use haiku by default, allowing appropriate model selection based on reasoning depth required for each task in the backlog.

What happens when verification fails for a task in automated execution?

The Skill loops continuously until the backlog is empty, enforcing verification rules strictly; tasks failing Playwright checks remain unresolved until verification passes, preventing incomplete work from moving to done status.