Why shouldn't I start with parallel execution immediately?

Starting manually helps discover naming conventions, content patterns, and potential errors that are not obvious at first glance, preventing those errors from being replicated across the entire batch.

How does the subagent prompt template work?

The template helps you document your manual findings into a structured instruction set for Claude, including transformation examples and skip conditions to guide parallel subagents.

What kind of tasks benefit most from this skill?

Any task involving many similar items, such as migrating skill files, restructuring components, renaming files across a directory, or batch cleaning data.

What is the Learn-First Parallelize pattern?

It is a meta-workflow where you perform 2-3 tasks manually to identify patterns and edge cases before using subagents to process the remaining items in parallel.

Learn-First Parallelize Pattern

Name: Learn-First Parallelize Pattern
Author: mshuffett

bymshuffett

0•

생산성 및 워크플로

Optimizes batch operations by discovering patterns manually before scaling with parallel subagents for high-efficiency refactoring.

The Learn-First Parallelize pattern provides a structured meta-workflow for handling large-scale migrations, refactorings, or data processing tasks with high confidence. By performing a small number of tasks manually first, the skill identifies naming conventions, edge cases, and quality criteria that might be missed in a fully automated approach. These learnings are then codified into a comprehensive subagent prompt, allowing Claude to launch parallel processes that maintain consistent quality across hundreds of items.

주요 기능

01Early identification of edge cases to prevent scaled errors

02Standardized subagent prompt template for repeatable task delegation

03Structured transformation examples for high-fidelity subagent performance

04Three-phase workflow: Manual Discovery, Documentation, and Parallel Execution

050 GitHub stars

06Quality checklists to ensure consistent output across parallel batches

사용 사례

01Performing mass refactors across large repositories where patterns vary slightly

02Migrating dozens of legacy code components to a new framework or library

03Batch processing configuration files or documentation with standardized cleanup

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add mshuffett/dotfiles learn-first-parallelize

For use in Claude.ai and ChatGPT

Download Skill