GoYoke banner
Bucket-Chemist Bucket-Chemist

GoYoke

Development community

Description

goYoke - Multi-agent orchestration framework for Claude Code

Installation

This entry records only its repository, not the path inside it, so there is no exact command to give. Open the source below and copy the folder into ~/.claude/skills/, or the file into ~/.claude/agents/.

README

goYoke

[![Build](https://github.com/Bucket-Chemist/goYoke/actions/workflows/build-test.yml/badge.svg)](https://github.com/Bucket-Chemist/goYoke/actions/workflows/build-test.yml) [![Go](https://img.shields.io/github/go-mod/go-version/Bucket-Chemist/goYoke)](https://go.dev/) [![License: MIT](https://img.shields.io/badge/License-MIT-blue.svg)](LICENSE) [![Platforms](https://img.shields.io/badge/platforms-linux%20%7C%20macOS%20%7C%20windows-brightgreen)]()

**Plan first. Route deliberately. Ship reviewed code.**

goYoke is a single Go binary that wraps Claude Code with runtime enforcement, typed agent contracts, multi-agent workflows, and a terminal UI for tracking sessions, dispatch, progress, and cost.

It is built for people who use AI coding tools for real software work and want less chat drift, fewer unreviewed one-shot changes, and clearer evidence that work was planned, delegated, checked, and completed. That premise lines up with recent findings that experienced developers value agents as productivity tools, but retain control over design and implementation because software quality still matters [1].

Built for Claude Code at launch. Multi-provider support is planned.

Why goYoke?

Claude Code is powerful, but raw chat workflows are easy to misuse:

  • exploratory work turns into implementation before the scope is known
  • expensive frontier models get used for simple search or boilerplate
  • reviewers produce prose that is hard to compare between runs
  • long sessions accumulate context noise and hidden assumptions
  • handoffs between "planner", "reviewer", and "worker" roles are informal

goYoke puts a structured harness around that work. It encourages a scout -> plan -> review -> implement loop, routes tasks to appropriate model tiers, and uses JSON-schema-validated contracts so agents produce comparable outputs instead of freeform theatre. This follows the same basic architectural direction as production agent guidance: prefer simple, composable workflows