ORCH — CLI orchestrator for AI coding agents, built on Node 20+ ESM + Commander.js

[oxgeneral]

Hey Hono community!

Love Hono for lightweight APIs. I built ORCH — a CLI runtime for orchestrating AI coding agent teams, using Node 20+ ESM with Commander.js (similar philosophy: do one thing well, no bloat).

What ORCH does:

npm install -g @oxgeneral/orch
orch init

orch agent add "Dev" --adapter claude     # Claude Code
orch agent add "Dev2" --adapter codex     # OpenAI Codex  
orch agent add "QA" --adapter claude      # Reviewer

orch task add "Build REST API" -p 1       # High priority
orch task add "Write tests" -p 2          # Medium priority

orch tui  # Live Ink/React TUI dashboard

Technical details Hono users might like:

• ESM-native: "type": "module", all imports use .js extension
• Zero frameworks in the core engine — layered DDD architecture
• Commander.js for CLI (thin outer layer)
• Ink/React for TUI (separate from business logic)
• Dependency injection via simple container (no decorators, no IoC frameworks)
• All state in YAML/JSON/JSONL files — no database

State machine:
todo -> in_progress -> review -> done (enforced, no shortcuts)

GitHub: https://github.com/oxgeneral/ORCH
npm install -g @oxgeneral/orch

1493 Vitest tests, TypeScript strict, MIT.

Node/TS folks — how are you managing multiple AI agent processes in your projects? Would love to compare approaches!