Software Engineering Collective

Tools / 03 · SI · Programming agent

Omega

Status: Open source · Pre-release

Omega is an open-source Python programming agent, published as Omes Bot, that consolidates the Hermes Agent runtime and the oh-my-pi agent harness into one programming interface. Developers configure a single-seat desk from a skills folder, a tool roster and a prompt folder. Built by Spectrum Web Co in Sydney.

Omega gives a developer a single seat for agentic programming: your skills, your tools and your prompts, driven by a runtime and harness that no longer need to be wired together by hand. The whole desk is defined in files you can read, version and review.

View the repo on GitHubTalk to us about agentic delivery

About

What is Omega?

Omega consolidates two agent projects into one programming interface, so a developer configures a desk once, with skills, tools and prompts, instead of maintaining separate setups and the glue between them. It is published on GitHub as Omes Bot.

  • Hermes Agent runtime plus oh-my-pi harness
  • A desk defined by skills, tools and prompts
  • Modular Python, open on GitHub

01 — One desk

Two projects, one seat.

The Hermes Agent runtime and the oh-my-pi harness sit behind a single programming interface. The codebase is modular, and upstream code is credited with its licences in the repository’s vendor notes and kept out of the runtime path.

  • Single-seat design
  • Upstream credit and licences in the repo
  • Modular Python codebase

02 — Configure

Skills, tools, prompts.

Point Omega at a skills folder, a tool roster and a prompt folder. The desk is defined by plain files you can read end to end, version alongside your code, review in a pull request and share with your team.

  • Skills folder
  • Tool roster
  • Prompt folder

03 — Run a task

One seat, start to finish.

A programming task starts at the seat and runs through the Hermes Agent runtime and the oh-my-pi harness behind one interface, drawing on the skills, tools and prompts you configured. You work from one place instead of switching between two agent setups, and the behaviour of the desk is defined by files your team can review.

  • One interface for two agent projects
  • Your skills, tools and prompts on every task
  • Configuration you can review in a pull request

04 — Status

Active development.

Omega is under active development with no formal releases yet, and the roadmap is tracked in the repository. Expect interfaces to change, and pin a known commit for anything you depend on.

Who it’s for

For developers who want one clean desk.

Omega is for developers who want a single, configurable programming agent they can read end to end, with its behaviour defined in files rather than hidden in a dashboard.

  • Python developers building with agents
  • Hermes Agent and oh-my-pi users
  • Engineers who prefer files to dashboards

FAQ

Omega, answered.

01

How does Omega combine Hermes Agent and oh-my-pi?

Omega consolidates the Hermes Agent runtime and the oh-my-pi agent harness into one Python programming interface. Instead of wiring a runtime and a harness together by hand, a developer works with a single desk that already brings them together. The codebase is modular Python, upstream code is credited in the repository’s vendor notes and kept out of the runtime path, and the whole project is published openly on GitHub under the name Omes Bot.

02

How do I configure Omega with skills, tools and prompts?

Omega is configured by pointing it at three things: a skills folder, a tool roster and a prompt folder. Together they define the desk, meaning what the agent knows how to do, which tools it is allowed to call and how it is instructed. Because the configuration is plain files rather than settings buried in a dashboard, a desk can be read end to end, versioned alongside your code, reviewed in a pull request and shared across a team. Once the desk is configured, you give Omega programming tasks from that single seat.

03

Is Omega (Omes Bot) production-ready?

Omega is in active development and has no formal releases yet, so it should be treated as a working tool rather than a finished product. The code is open on GitHub, the roadmap is tracked in the repository, and interfaces may change between commits. Teams evaluating it for serious work should pin a known commit, review changes before upgrading and keep the tool roster limited to what each task needs. That keeps an early-stage agent predictable while the project matures, and the repository is the place to follow its progress.

04

How does Omega credit upstream open-source projects?

Omega credits every upstream project it builds on in the repository’s vendor notes, together with the licences that apply to that code. The Hermes Agent runtime and the oh-my-pi harness are both upstream projects, and keeping their provenance explicit lets anyone auditing the codebase see where each part came from and under which terms. Upstream code is kept out of the runtime path, which keeps Omega’s own modules distinct from what it vendors. For organisations that run open-source licence reviews, the vendor notes are the place to start.

05

Should I use Omega or Agent Swarm?

Choose Omega when one developer wants a single, configurable programming agent, and Agent Swarm when a team needs many agents coordinated across the software lifecycle. Omega is a single-seat desk: one agent, driven by your skills, tools and prompts, built on the Hermes Agent runtime and the oh-my-pi harness. Agent Swarm is a hierarchy of 15 specialised agents under an orchestrator, with signed handoffs, fail-closed gates and human escalation. Both are open source from Spectrum Web Co, so a team can start with Omega for individual work and evaluate Agent Swarm for delivery at team scale.