sadd:do-in-parallel

Coordinate parallel sub-agents across files with model selection and self-critique.

Updated Mar 4, 2026
One-click install
npx skills add https://github.com/dalawwa/labor-methods --skill sadd-do-in-parallel
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: sadd:do-in-parallel
Source: https://github.com/dalawwa/labor-methods/tree/main/.cek/plugins/sadd/skills/do-in-parallel
Command: npx skills add https://github.com/dalawwa/labor-methods --skill sadd-do-in-parallel

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Launching many tasks across files or targets often leads to wasted time and inconsistent results. This Skill orchestrates parallel sub-agents, automatically selecting the best model and applying quality-focused prompting to accelerate batch work while maintaining output quality.

Core Features & Use Cases

  • Parallel task dispatch across files or named targets for batch processing of large codebases or data sets
  • Unified model selection strategy (opus/sonnet/haiku) to balance quality and cost
  • Zero-shot Chain-of-Thought reasoning with mandatory self-critique to improve reliability
  • Phase-driven workflow: target identification, task analysis, per-target prompts, self-critique, parallel dispatch, and result collection
  • Independence and safety checks to ensure targets can run in parallel without shared state conflicts

Quick Start

Supply a task description and target sources (files or named targets) to initiate parallel sub-agent execution.

Frequently Asked Questions about sadd:do-in-parallel

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

FAQPage Schema
How do I run batch tasks across multiple files in parallel?

To run batch tasks in parallel, you dispatch sub-agents across files or named targets. This Skill coordinates execution by applying deterministic per-target prompts and collecting structured outputs automatically.

What is the best way to analyze multiple modules in parallel?

Analyzing multiple modules in parallel is best handled by launching sub-agents with a unified model selection strategy. The system balances cost using opus, sonnet, or haiku models while maintaining independence checks.

How do I ensure consistent results when applying the same transformation across many files?

To ensure consistent results across many files, parallel orchestration applies zero-shot Chain-of-Thought reasoning with mandatory self-critique. This quality-focused prompting enforces deterministic outputs per target.

Can I generate documentation for multiple components in parallel without state conflicts?

Yes, you can generate documentation for multiple components in parallel safely. The workflow includes built-in independence checks to ensure targets execute without shared state conflicts during dispatch.

Do I need to manually select models for parallel sub-agent dispatch?

No, you do not need to manually select models for parallel sub-agent dispatch. The orchestration applies a unified model selection strategy that automatically balances cost across opus, sonnet, and haiku models.

What are the limitations of using sub-agents for parallel task dispatch?

A key limitation of parallel task dispatch is the requirement for target independence. The system performs safety checks to prevent shared state conflicts, meaning tasks dependent on prior outputs cannot run simultaneously.