Omnigent

Working

A common layer over Claude Code, Codex and the agents you write yourself

10.0k GitHub starsPushed 1 day agoInstall on your machine1 checkDeveloper utilities

Open Omnigent SourceSite checked 2 days ago

What stands out

Taken from the page and the access facts we recorded. Nothing here is guessed.

put spend caps and policies around agents you already run
work on the same live agent session with someone else
swap the harness underneath without rewriting your agents
sandbox an agent whose blast radius you are unsure of
Free
Uses your own API key
Source is public

What we checked

A missing tick means we have not measured it. A red mark is a check that failed.

App accessible
The project site answered. That is not the installed tool running.
Page rendered
No browser probe yet. An HTTP 200 is not this check.

Looked: 0
Opened: 1
Checks: 1
Watching: 0

Looked and opened count distinct browsers on this site, not verified people. Checks are requests we made. None of these come from the app itself.

Screenshot of Omnigent — Screenshot from our last capture

Development

3 snapshots kept

Active

Commits landed in the last 90 days.

+31 stars since we started watching.

Evidence

Each bar is one request to the project's own site

This one is installed and run on your own machine, so what we re-check is the project’s site — whether it is still published and still reachable. We cannot tell you from here that the tool itself runs.

Checks: 1
Working: 100%
Median response: 191ms
Watching since: 2 days ago

Activity

Watching since 2 days ago

Last content change: None seen yet
Unchanged for: 2 days
Certificate expires: in 67 days

How it looked

11 hours ago

We hash the visible text on every check, so a redesign registers and a rotating advert does not. An app that has not changed in a long time is not necessarily abandoned — but it is a fact worth knowing before you depend on it.

What it does

An open-source meta-harness for composing and governing AI agents on top of harnesses you already use, so you can swap or combine them without rewriting. It adds contextual policies covering spend caps, model routing and risk-based escalation, plus sandboxing, and real-time collaboration on the same live session from any device. Agents are defined in YAML, with a coding orchestrator and a model-debate agent built in.

What people say

No reviews yet.

We would rather show none than invent them.

Embed this status

Put it in your README. It updates itself every time we check, and links to your verification record — evidence a stranger can read, not a page selling them something.

[![Omnigent — still works](https://itstillworks.cruscible.com/api/badge/omnigent)](https://itstillworks.cruscible.com/verify/omnigent)

Others like this

Each of these answered on its most recent check.