[R0] Verify one plain OpenClaw deployment on NemoClaw v1 through Brev CLI

## Outcome

One pinned v1 revision provisions a standard OpenClaw agent on the selected Brev host and produces a real model-backed answer.

## Scope and ownership

Can proceed alongside main configuration generation and shared-contract work. This is ordinary deployment verification, not a custom agent or a voice simulation. No full main-to-v1 migration or every managed backend is required.

Proposed owner: NemoClaw v1 deployment/integration owner; individual maintainer to be named.

This ticket records requested work. New product boundaries and supported combinations require the owning maintainer decision; no existing proposal is declared accepted here.

## Work

- Record the agreed Brev host/profile, current v1 revision, runtime bundle, OpenShell version, agent image and inference dependencies.
- Use Brev CLI to access the selected test host and run the actual v1 deployment path. An external OpenShell gateway and external inference are acceptable documented prerequisites for the initial supported profile.
- Verify native-agent readiness and one real answer using ordinary authorized OpenClaw access.
- Reapply the unchanged configuration and verify the existing agent/deployment identity is preserved.
- Record unresolved v1 provisioning/runtime gaps without substituting main or the older v1-poc implementation.

## Acceptance

- [ ] Commands and environment are reproducible through Brev CLI.
- [ ] The deployed revision is explicitly v1 and the agent/model are real.
- [ ] Repeated unchanged apply preserves the established deployment.
- [ ] Evidence distinguishes native-agent verification from VoiceClaw attachment; this baseline alone does not complete R0.

## Dependencies

R0 parent epic: [#11746](https://github.com/NVIDIA/NemoClaw/issues/11746).

VoiceClaw counterpart: [VoiceClaw #7](https://gitlab-master.nvidia.com/jarvis/voice-claw/-/issues/7).

Can proceed in parallel with the other foundational work.

## Category

Testing

Start by confirming the agreed Brev host/profile, pinned NemoClaw v1 revision, runtime bundle, OpenShell version, agent image, and inference dependencies. Use Brev CLI to run the actual v1 deployment path, then reapply the unchanged configuration. Done means reproducible commands, one real model-backed answer, preserved deployment identity, and separate evidence for native-agent verification versus VoiceClaw attachment.

Written by the indexing model from the issue text.