Operator governance workbench adjacent to OpenShell

[Epydios]

When I saw OpenShell, I saw a lot of similarities to some of the things I have been building.

Therefore, I want to share my OSS project that seems to be pretty complementary to OpenShell:
https://github.com/Epydios/Epydios-Agent-Operations-Plane

As opposed to sandbox, it is focused more on the operator side of governed execution:

• review and approval flow on top of execution
• audit and evidence continuity attached to the same run path
• versioned provider boundaries that include: policy, evidence, profile resolution and desktop action handling
• an operator desktop/workbench GUI

This repo is an example of the directions that I’ve been working on for a while. It doesn't sit in the exact same space as OpenShell, but definitely on the same playing field...a field that, to my thinking, is going to become pretty important as tolerance for agentic operational risk is reduced, especially in areas like finance, healthcare, government, etc.

I'd be curious to hear any thoughts and happy to discuss further.

[Advait251206]

This is an interesting direction, and I agree that it seems more complementary than competitive.

From your description, OpenShell and your project appear to operate at different layers of the stack:

• OpenShell focuses on providing a secure, isolated execution environment for agents.
• Epydios Agent Operations Plane focuses on the operational governance around those executions—approval workflows, audit trails, policy enforcement, evidence collection, and operator tooling.

Those concerns naturally fit together, especially in regulated environments where simply sandboxing execution isn't enough. Being able to answer questions like who approved this action, what policy was applied, what evidence was collected, and why was a particular decision made is just as important as isolating the execution itself.

The separation of concerns also seems appealing architecturally:

• OpenShell could remain responsible for sandbox lifecycle and execution.
• An operations/governance layer could orchestrate approvals, maintain audit history, and expose an operator-facing UI without coupling those concerns into the runtime itself.

If anything, it sounds like there could be opportunities for integration rather than overlap—for example, exposing sandbox lifecycle events or execution metadata that an external governance platform could consume.

I'd be interested to hear the maintainers' thoughts on whether they envision OpenShell remaining focused on execution infrastructure, with governance handled by external systems, or whether some of these operator-oriented capabilities are part of the long-term roadmap.

Thanks for sharing the project—it's always interesting to see complementary approaches emerging around the same ecosystem.

---

If this answer helped or pointed you in the right direction, I'd appreciate it if you could mark it as the accepted answer so it's easier for others with the same issue to find.

Also, if you found my contribution useful, I'd appreciate it if you could check out my GitHub profile, follow me, and star any repositories you find interesting.

GitHub: https://github.com/Advait251206