agents-in-a-box

   ╔═══════════════════════════════════════════════════════════════╗
   ║                                                               ║
   ║     █████╗  ██████╗ ███████╗███╗   ██╗████████╗███████╗       ║
   ║    ██╔══██╗██╔════╝ ██╔════╝████╗  ██║╚══██╔══╝██╔════╝       ║
   ║    ███████║██║  ███╗█████╗  ██╔██╗ ██║   ██║   ███████╗       ║
   ║    ██╔══██║██║   ██║██╔══╝  ██║╚██╗██║   ██║   ╚════██║       ║
   ║    ██║  ██║╚██████╔╝███████╗██║ ╚████║   ██║   ███████║       ║
   ║    ╚═╝  ╚═╝ ╚═════╝ ╚══════╝╚═╝  ╚═══╝   ╚═╝   ╚══════╝       ║
   ║              ██╗███╗   ██╗    █████╗                              ║
   ║              ██║████╗  ██║   ██╔══██╗                             ║
   ║              ██║██╔██╗ ██║   ███████║                             ║
   ║              ██║██║╚██╗██║   ██╔══██║                             ║
   ║              ██║██║ ╚████║   ██║  ██║                             ║
   ║              ╚═╝╚═╝  ╚═══╝   ╚═╝  ╚═╝                             ║
   ║            ██████╗  ██████╗ ██╗  ██╗                              ║
   ║            ██╔══██╗██╔═══██╗╚██╗██╔╝                              ║
   ║            ██████╔╝██║   ██║ ╚███╔╝                               ║
   ║            ██╔══██╗██║   ██║ ██╔██╗                               ║
   ║            ██████╔╝╚██████╔╝██╔╝ ██╗                              ║
   ║            ╚═════╝  ╚═════╝ ╚═╝  ╚═╝                              ║
   ║                                                               ║
   ╚═══════════════════════════════════════════════════════════════╝

A complete ecosystem for AI-assisted development

115 Rust Modules · 71 Skills · 37 Agents · 9 AI Tools · Knowledge Graph

A terminal-native ecosystem for managing AI coding agents. Built around a Rust TUI that orchestrates Claude Code sessions with git worktree isolation, and a portable toolkit of skills, agents, and workflows that plug into 9 different AI coding tools.

Creating a new Claude Code session with model selection

What's Inside

Why agents-in-a-box?

Most AI coding setups are a loose collection of dotfiles. This project treats the problem as an engineering system:

• One toolkit, many tools — Write a skill once, deploy it to Claude Code, Codex, Gemini, Cursor, Copilot, Amazon Q, Cline, Roo, or Clawdhub
• Session isolation — Each coding session gets its own git worktree and tmux session. No cross-contamination
• Agents that compose — 37 specialized agents (backend-developer, security-agent, architecture-reviewer, etc.) that can be orchestrated into swarms
• Memory that persists — A two-tier knowledge system (GraphRAG + QMD) that captures learnings and retrieves them across sessions and projects
• Production Rust — The TUI isn't a shell script. It's 115 modules of typed, tested, async Rust with clippy pedantic/nursery lints

Quick Start

# Install the TUI
brew tap stevengonsalvez/ainb && brew install ainb

# Install the toolkit for your AI tool
cd toolkit && npm install && node create-rule.js --tool=claude-code-4.5

# Launch
ainb

ainb — Terminal UI

A Rust-based terminal application for managing Claude Code development sessions with git worktree isolation, model selection, and persistent tmux sessions.

Features

• Session management — Create, monitor, switch between isolated Claude Code sessions
• Git worktree isolation — Each session gets its own branch and working directory
• Model selection — Choose Sonnet, Opus, or Haiku per session
• Live log streaming — Real-time log viewer with level filtering and search
• tmux integration — Persistent sessions that survive disconnects
• Keyboard-driven — Vim-style navigation throughout

Main dashboard with active sessions	Session creation with model selection
Real-time log streaming	Attached Claude Code session

Installation

brew tap stevengonsalvez/ainb
brew install ainb

curl -fsSL https://raw.githubusercontent.com/stevengonsalvez/agents-in-a-box/v2/ainb-tui/install.sh | bash

cargo install --git https://github.com/stevengonsalvez/agents-in-a-box --branch v2 agents-box
# Optionally alias: alias ainb="agents-box"

# 1. Install WSL2
wsl --install

# 2. Inside Ubuntu/Debian
curl -fsSL https://raw.githubusercontent.com/stevengonsalvez/agents-in-a-box/v2/ainb-tui/install.sh | bash
sudo apt update && sudo apt install -y tmux
ainb

ainb requires tmux for persistent sessions, which is Unix-only. WSL provides the best Windows experience.

Keyboard Shortcuts

Platform Support

Requirements

• tmux — persistent session management
• git — worktree operations
• Claude Code CLI — the claude command

Toolkit

A portable AI coding agent toolkit: skills, agents, workflows, and configurations that deploy to 9 different AI coding tools from a single source.

Full toolkit documentation →

Supported AI Tools

Skills (71)

Skills are reusable capabilities that any supported AI tool can invoke.

plan · plan-tdd · plan-gh · implement · validate · workflow · brainstorm · critique · discuss · expose · interview

commit · find-missing-tests · webapp-testing · security-audit · security-scan · simplify

start-local · start-ios · start-android · spawn-agent · tmux-monitor · tmux-status · expose · debug-bridge

reflect · global-learnings · research · research-cache · instincts · compound-docs · prime

health-check · session-info · session-metrics · session-summary · handover · recover-sessions · plugins

swarm-create · swarm-join · swarm-inbox · swarm-status · swarm-shutdown · swarm-orchestration · swarm-agent-troubleshooting

gh-issue · make-github-issues · do-issues · merge-agent-work · list-agent-worktrees · attach-agent-worktree · cleanup-agent-worktree

ui-ux-pro-max · frontend-design · frontend-slides · tui-style-guide · tui-screen · liquid-glass · remotion-best-practices

crypto-research · oracle · notebooklm · sentry-cli · ats-resume-matcher · resume-formatter · retro-pdf

skill-creator · agent-ops · autonomous-loops · cost-aware-pipeline · media-processing · nano-banana-pro · sync-learnings · claude-developer-platform

Agents (37)

Specialized AI agents organized by domain. Each agent has a defined persona, tool access, and area of expertise.

Knowledge System

A two-tier learning system that captures insights during development and retrieves them across sessions and projects.

The /reflect skill captures learnings. The /research and /prime skills retrieve them. The /global-learnings skill manages the knowledge base directly.

How the knowledge system works →

Architecture

agents-in-a-box/
│
├── ainb-tui/                   # Rust TUI application
│   ├── src/                    # 115 modules
│   │   ├── app/                #   State machine & event handling
│   │   ├── components/         #   TUI screen components
│   │   ├── widgets/            #   Reusable UI widgets
│   │   ├── docker/             #   Container management
│   │   ├── tmux/               #   Session & PTY integration
│   │   ├── git/                #   Worktree operations
│   │   ├── claude/             #   Claude API client
│   │   ├── models/             #   Data models
│   │   └── config/             #   Configuration handling
│   ├── deny.toml               #   License & security policy
│   ├── Formula/                #   Homebrew formula
│   └── install.sh              #   One-liner installer
│
├── toolkit/                    # Portable AI agent toolkit
│   ├── packages/
│   │   ├── skills/             #   71 reusable skills
│   │   ├── agents/             #   37 agent definitions
│   │   │   ├── universal/      #     Cross-stack specialists
│   │   │   ├── engineering/    #     Backend & infra agents
│   │   │   ├── orchestrators/  #     Team coordination
│   │   │   ├── design/         #     UI/UX specialists
│   │   │   ├── swarm/          #     Multi-agent coordination
│   │   │   └── meta/           #     Agent creation & reflection
│   │   ├── workflows/          #   Structured delivery workflows
│   │   └── utilities/          #   Shared utilities
│   ├── bootstrap.js            #   Multi-tool deployment engine
│   └── create-rule.js          #   CLI installer
│
├── docs/                       # Documentation
│   └── how-reflection-works.md #   Knowledge system architecture
│
└── .github/workflows/
    ├── ci.yml                  #   Rust CI (fmt, clippy, test, deny, machete)
    ├── toolkit-validation.yml  #   Toolkit structure & install validation
    └── release.yml             #   Cross-platform binary releases

CI/CD & Quality

The Rust codebase enforces unsafe_code = "forbid" and runs clippy with pedantic, nursery, and cargo lint groups enabled.

Development

Building from source

cd ainb-tui
cargo build --release
./target/release/agents-box

Running tests

cd ainb-tui
cargo test                              # Unit tests
cargo test --features visual-debug      # With terminal output
cargo test --features vt100-tests       # VT100 screen verification
cargo nextest run                       # With nextest (parallel)

Linting & checks

cd ainb-tui
cargo fmt --check                       # Format check
cargo clippy --all-targets              # Lint
cargo deny check                        # Security + licenses

Installing the toolkit

cd toolkit
npm install
node create-rule.js --tool=claude-code-4.5    # Deploy to ~/.claude/
node create-rule.js --tool=gemini             # Deploy to .gemini/
node create-rule.js --tool=codex              # Deploy to ~/.codex/

Contributing

1. Fork the repository
2. Create a feature branch (git checkout -b feature/amazing-feature)
3. Commit your changes (git commit -m 'feat: add amazing feature')
4. Push to the branch (git push origin feature/amazing-feature)
5. Open a Pull Request

Links

• Releases
• Homebrew Tap
• Issues
• Knowledge System Architecture
• Toolkit Documentation

License

MIT — see LICENSE for details.