Which local tools are required for this skill?

This skill requires Git (for worktree management) and mprocs (to handle the concurrent execution of multiple AI workers).

Where does the skill store its session data?

All session metadata, including the tasks.json and worker prompts, is stored within the .hive/sessions directory in your project root.

How are the three implementations different?

Worker A focuses on clean and readable code, Worker B explores alternative logic or data structures, and Worker C prioritizes maximum performance.

What is a Fusion Thread or F-thread?

An F-thread is a development workflow where three distinct implementations of a single algorithm are generated in parallel worktrees and then evaluated to select the best one.

How does the 'Judge Queen' decide on the winner?

The Judge Queen compares all three versions based on implementation correctness, algorithmic complexity, code readability, and the results of performance tests.

Fusion Algorithm Optimizer

Name: Fusion Algorithm Optimizer
Author: rdfitted

byrdfitted

0•

開発者ツール

Optimizes algorithm selection by generating and comparing three distinct implementation approaches in parallel Git worktrees.

The Fusion Algorithm skill automates the process of identifying the best programmatic approach for a given problem by employing an 'F-thread' strategy. It orchestrates three separate AI workers in dedicated Git worktrees—focusing on readability, alternative logic, and raw performance respectively—before a 'Judge Queen' agent evaluates the results. This skill is ideal for complex logic problems where multiple valid architectural or algorithmic paths exist, ensuring you choose the most efficient, maintainable, and correct solution without manual trial and error.

主な機能

01Automated comparison and selection of the winning implementation by a judge agent

02Triple-worker strategy targeting readability, alternative logic, and performance

03Automated Git worktree management for parallel implementation development

04Multi-process orchestration using mprocs for concurrent execution

05Standardized session tracking and task logging via .hive metadata

060 GitHub stars

ユースケース

01Optimizing performance-critical backend logic by testing multiple data structures

02Resolving implementation uncertainty by exploring multiple architectural patterns simultaneously

03Refactoring legacy algorithms to improve readability without sacrificing speed

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add rdfitted/claude-code-setup fusion-algorithm

For use in Claude.ai and ChatGPT

Download Skill