Problem Statement
As a new user (of both NemoClaw and OpenClaw), I'm trying to configure OpenClaw to add Telegram groups to my already-onboarded agent (so they can respond to @ mentions), use my self-hosted SearXNG search engine etc.
AFAIK things like this all require modifying the openclaw.json, which is locked after setup. And the onboarding process currently does not even expose these, and/or they also might change after deployment.
Running
openclaw configure
from the sandbox gives
Error: 'openclaw configure' cannot modify config inside the sandbox.
The sandbox config is read-only (Landlock enforced) for security.
I understand this is expected/desired for nemoclaw. However, as a new user this is confusing/frustrating for several reasons:
- These are the instructions provided by the official OpenClaw documentation, and NemoClaw/the NemoClaw documentation provides no clear alternative path
- The onboarding process does not currently expose the majority of OpenClaw settings
- Even if it did, I might not be fully aware of every single configuration I need during onboarding.
Proposed Design
The ask here is to provide a way to update the openclaw.json (and other critical settings files) after installation:
- Internally from within the sandbox so OpenClaw's native configuration tools can be used (a mechanism to request/grant write access to openclaw.json via OpenShell?)
- Externally from the sandbox somehow?
- Provide a way to specify or manually configure any/all settings when installing/onboarding NemoClaw?
Any of these approaches would be a greatly appreciated improvement.
Alternatively, if this already exists, it would be great if this could be documented and included in the "Getting Started" / "Install NemoClaw and Onboard OpenClaw Agent" section of the README.md for new users to see
Alternatives Considered
As per the proposed design section :)
Category
enhancement: feature
Checklist
Problem Statement
As a new user (of both NemoClaw and OpenClaw), I'm trying to configure OpenClaw to add Telegram groups to my already-onboarded agent (so they can respond to @ mentions), use my self-hosted SearXNG search engine etc.
AFAIK things like this all require modifying the openclaw.json, which is locked after setup. And the onboarding process currently does not even expose these, and/or they also might change after deployment.
Running
openclaw configurefrom the sandbox gives
I understand this is expected/desired for nemoclaw. However, as a new user this is confusing/frustrating for several reasons:
Proposed Design
The ask here is to provide a way to update the openclaw.json (and other critical settings files) after installation:
Any of these approaches would be a greatly appreciated improvement.
Alternatively, if this already exists, it would be great if this could be documented and included in the "Getting Started" / "Install NemoClaw and Onboard OpenClaw Agent" section of the README.md for new users to see
Alternatives Considered
As per the proposed design section :)
Category
enhancement: feature
Checklist