ultrawork

Dispatch multiple independent agent calls concurrently with model-tier routing.

Updated Aug 23, 2026
One-click install
npx skills add https://github.com/kamelmh/opencode --skill ultrawork-kamelmh
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: ultrawork
Source: https://github.com/kamelmh/opencode/tree/main/skills/ultrawork
Command: npx skills add https://github.com/kamelmh/opencode --skill ultrawork-kamelmh

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Ultrawork solves slow, sequential work when multiple independent tasks could be executed at the same time, letting you complete more in less time.

Core Features & Use Cases

  • Parallel agent execution: Launches multiple independent agent tasks simultaneously to increase throughput.
  • Model-tier routing: Routes each subtask to the appropriate agent/model tier (low/medium/high) to balance quality and cost.
  • Composable execution layer: Designed to provide parallelism without persistence loops; higher-level modes (ralph/autopilot) add verification and lifecycle control.

Quick Start

Use ultrawork to run several independent implementation and testing tasks at once by delegating the plan to a parallel executor with the right model tiers.

Frequently Asked Questions about ultrawork

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

FAQPage Schema
How do I run multiple independent coding tasks in parallel to speed up development?

Parallel execution of independent coding tasks is achieved by dispatching multiple agent calls concurrently. This allows you to implement multiple features, generate tests, or run build-related tasks simultaneously rather than sequentially.

What is model routing and how does it work for parallel agent tasks?

Model routing assigns each parallel subtask to a specified agent or model tier, such as low, medium, or high. This mechanism balances processing quality against computational cost for every independent operation dispatched.

Can I execute long-running build commands in the background while making smaller edits?

Yes, long-running build commands are dispatched using background execution while smaller independent edits proceed. Lightweight completion checks are applied to verify when the background operations finish.

What is the best way to handle task delegation for multiple independent features?

Delegating independent features to a parallel executor with explicitly specified model tiers is the best approach. It fires operations simultaneously without persistence loops, increasing overall development throughput.

When should I not use parallel execution for software engineering tasks?

Parallel execution should not be used when tasks lack clear independence. It is designed strictly for scenarios with independent work items like generating tests or running build tasks alongside separate edits.