> great.sh

The managed AI dev environment. Alpha — open to testing and feedback.

One command. 15 AI agents. Fully configured. Open source.

Quick Start

curl -sS https://great.sh/install.sh | sh
great init
claude

What it does

• One command setup — from a blank machine to a fully configured AI dev environment
• AI agent orchestration — 15 specialized agents installed into Claude Code with great loop install
• MCP server management — install, configure, and credential-inject MCP servers from the registry
• Cloud-synced credentials — zero-knowledge encrypted vault syncs API keys across machines

CLI Commands

Command	Description
great init	Initialize a new great.sh environment
great apply	Apply configuration to the current environment
great status	Show environment status
great sync	Sync configuration with the cloud
great vault	Manage encrypted credentials
great mcp	Manage MCP servers
great doctor	Diagnose environment issues
great update	Update great.sh to the latest version
great diff	Show configuration diff
great template	Manage configuration templates
great loop	Install and manage the great.sh Loop agent team

MCP Bridge

The great mcp-bridge command runs an MCP server that routes to multiple AI backends (Gemini, Codex, Claude, Grok, Ollama).

Add it to Claude Code for the current project:

claude mcp add great-bridge -- great mcp-bridge

Or add it globally (all projects):

claude mcp add --scope user great-bridge -- great mcp-bridge

The great.sh Loop

A 16-role AI agent orchestration methodology (4 teammates + 11 subagents + 1 team lead), shipped as a Claude Code plugin. Each role is embodied by a historical figure whose expertise maps to the task.

Nightingale → Lovelace → Socrates → Humboldt → Da Vinci →
  Wirth / Turing / Kerckhoffs / Nielsen (parallel) →
  Dijkstra (code review) → Rams (visual) →
  Knuth + Gutenberg (docs) →
Hopper → Deming

Requirements → spec → review → scout → build → test/security/perf/UX → code review → visual → docs → commit → observe. One iteration at a time. Von Braun handles release builds separately via /great:deploy.

Each role runs on the model tier its job demands: Fable for the builder and the plan-approval gate, Opus for testing, security audit, specs, and code review (security roles stay on Opus deliberately — Fable's cyber safety classifiers can refuse security-focused analysis), Sonnet for scouting, UX, visual, perf, docs, and deploys, Haiku for commits. Deming, the team lead, is your own session — run it on Opus or Fable.

great loop install            # registers the plugin with Claude Code
great loop install --project  # adds .tasks/ working state to the current repo
claude                        # then: /great:backlog → /great:loop

Templates

Pre-configured environment templates from architecton.ai:

• AI Full Stack (TypeScript) — Claude Code + Codex + Gemini, Playwright MCP server, gh CLI
• AI Full Stack (Python) — Python with uv, PostgreSQL MCP server, full AI agent setup
• AI Data Science — CUDA, Jupyter, Gemini for data analysis, database MCP servers
• AI DevOps — Terraform, AWS CLI, Docker, Kubernetes MCP servers
• AI Minimal — Just Claude Code with Filesystem and Memory MCP servers

Development

# CLI
cargo build
cargo test
cargo clippy

# Marketing site
pnpm dev
pnpm build:site

Status

Alpha (v0.1.0). Core features work. We welcome bug reports and feedback.

• GitHub Issues
• Discussions

License

Apache-2.0 — see LICENSE

Copyright 2025 Superstruct Ltd