do:work

Execute work plans by dispatching steps to subagents with task tracking.

Updated Aug 23, 2026

Search Tags:#automation #workflow #task management #development workflow #agent orchestration #code execution

Ready to run

npx skills add https://github.com/adamfeldman/compound-workflows --skill do-work-adamfeldman

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: do:work
Source: https://github.com/adamfeldman/compound-workflows/tree/main/plugins/compound-workflows/skills/do-work
Command: npx skills add https://github.com/adamfeldman/compound-workflows --skill do-work-adamfeldman

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) components.

What problem does it solve?

This Skill automates the execution of complex work plans by breaking them down into manageable steps and delegating each step to an independent subagent, preventing context exhaustion and ensuring progress tracking.

Core Features & Use Cases

Subagent Architecture: Orchestrates multiple AI agents to execute distinct tasks within a larger plan.
Robust Task Tracking: Leverages tools like beads or TodoWrite for persistent progress tracking, surviving session compactions.
Worktree Isolation: Creates isolated development environments to prevent interference with the main codebase.
Use Case: Imagine a complex software feature requiring planning, coding, testing, and review. This Skill can dispatch each of these phases to specialized subagents, ensuring each part is handled correctly and progress is maintained across potentially long or interrupted sessions.

Quick Start

Use the do:work skill to execute the plan located at 'docs/plans/2023-01-15-feat-new-api-plan.md'.

Frequently Asked Questions about do:work

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

FAQPage Schema

How do I automate complex development workflows with subagent orchestration?▼

Subagent orchestration automates complex development workflows by dispatching individual plan steps to independent subagents. This approach prevents context exhaustion and ensures persistent progress tracking across potentially long or interrupted coding sessions.

What is the best way to manage isolated development environments for automated task execution?▼

Managing isolated development environments for automated task execution is best handled through automated worktree creation. This isolates subagent code execution from the main codebase, preventing interference during complex feature development.

How does task tracking survive session compactions during long automated code execution?▼

Task tracking survives session compactions during long automated code execution by leveraging persistent tracking systems like beads or TodoWrite. These tools maintain progress state across sessions, ensuring no steps are lost.

Can I use this subagent dispatch system to handle plan clarification and PR creation?▼

Yes, you can use this subagent dispatch system for plan clarification and PR creation. It handles subagent prompt construction, result review, and final shipping procedures including automated cleanup and pull request creation.

Do I need to manually create worktrees before dispatching tasks to subagents?▼

No, you do not need to manually create worktrees before dispatching tasks. The system features automated worktree creation and management, providing isolated development environments for each subagent without manual setup.