Introducing Omnigent: A Meta-Harness for Your Agents

Matei Zaharia, Kasey Uhlenhuth and Corey Zumar argue that coding agents now need a layer above the harness. Their users run Claude Code, Codex and Pi side by side, and every harness has its own session model, its own permissions and its own way of being shared, so nothing composes. Omnigent is Databricks’ answer: a common API over the lot of them, open sourced under Apache 2.0 and still alpha at the time of writing.

The abstraction rests on a claim that all these tools have the same shape underneath, messages and files in, text streams and tool calls out. Wrap each agent in a sandboxed session behind that uniform interface and you can add the things individual harnesses do not offer: session sharing by URL, web and mobile and API access to the same run, cloud execution, per-session spend limits, and security policies that track state rather than just allowing or denying calls.

The framing worth keeping is about where to invest. The models and harnesses will keep changing as the field evolves; the layer you work at shouldn’t have to. They also describe how their heaviest users have moved on from single-agent prompting: instead of prompting one agent at a time, they design loops that drive whole teams of agents. Bear in mind this is a vendor pitching its own new abstraction, and the post carries no adoption or benchmark numbers to back it, only the scale of the engineering org it came from.