omnigent vs squad
Open-source meta-harness over Claude Code, Codex, Cursor, OpenCode, Hermes, Pi and custom agents: swap harnesses without rewriting, enforce policies and sandboxes, follow sessions from any device. — versus — Multi-agent terminal collaboration for AI CLIs: a manager, workers and an inspector — Claude Code, Gemini, Codex, OpenCode — coordinating through one-shot shell commands and SQLite. No daemon.
Same job — coordinating multiple AI coding CLIs — opposite philosophy: squad is daemonless one-shot shell + SQLite, Omnigent is a full server/meta-harness.
| omnigent | squad | |
|---|---|---|
| Stars | 7.5k | 606 |
| Forks | 1.1k | 60 |
| Language | Python | Rust |
| License | Apache-2.0 | MIT |
| Last activity | today | 1 months ago |
| Topics | agents, orchestration, coding | coding, orchestration |
| Curated connections | 5 | 4 |
omnigent — the curator's take
Use when you run several coding harnesses and want one control plane: mix agents in a session, gate risky actions with policies, run in Modal/Daytona/E2B/K8s sandboxes, and pick a session up from phone or browser. Team sharing (co-drive, fork) is a real differentiator. NOT if you live in one CLI on one laptop — it's a server + desktop app stack, and it's alpha; expect churn. Apache-2.0.
squad — the curator's take
The unix-philosophy take on multi-agent coding: no daemon, no server, no framework — each agent is a terminal running the CLI you already use, coordinating through /squad slash commands backed by SQLite. Assign a manager, spin workers, add an inspector, watch them divide the work. Radically simpler to reason about than orchestration platforms, and it dies clean (every command is one-shot). NOT for production pipelines or unattended fleets — it's built for a human watching terminals; young (~600 stars) and quiet for weeks at review time, so kick the tires before making it a habit.