[R0] Demonstrate a v1-deployed agent reached through onboarded VoiceClaw using Brev CLI

## Outcome

Using Brev CLI, prove that an independent client reaches one v1-deployed OpenClaw agent through VoiceClaw with direct, scoped authentication.

This is the user-selected temporary R0 experiment. It preserves the ownership boundary in `VOICECLAW_NEMOCLAW_API_CONTRACT.md` but deliberately replaces its offer/redemption/attachment-lease mechanism with direct scoped authentication. It does not claim full `v1alpha1` conformance.

## Test steps

1. Apply hand-authored configuration directly in v1 on the selected Brev host. Verify the agent. Check that import/validation alone does not deploy or start onboarding.
2. Start with VoiceClaw absent. After agent readiness, v1 apply invokes VoiceClaw-owned onboarding. Observe package installation, prompts for its settings/credentials, configuration, startup and client connection instructions.
3. Issue one short-lived credential scoped to that agent. Hand VoiceClaw the private semantic API endpoint and credential through protected storage or a pipe.
4. Connect an independent client through VoiceClaw's real client API/SDK and transport. VoiceClaw authenticates directly to NemoClaw and confirms actual agent readiness.
5. Check invalid/expired credentials, wrong/replaced targets and unavailable native agent using focused fixtures or separate operator-controlled runs.
6. Stop and clean up the experiment through each resource owner. If repeated, issue a fresh credential manually through the operator workflow.

## Acceptance

- [ ] VoiceClaw owns its installation, onboarding and external process lifecycle.
- [ ] The client reaches the intended real agent through VoiceClaw, and VoiceClaw uses only NemoClaw's semantic API.
- [ ] Native protocols, IDs and credentials stay private to NemoClaw/OpenShell.
- [ ] SSH access, an open port or HTTP health alone does not count as client-to-agent reachability.
- [ ] The endpoint/credential handoff is protected; authentication, target binding, expiry and readiness checks pass.
- [ ] Record component revisions, commands and redacted evidence sufficient to repeat the run.
- [ ] No offer redemption, credential exchange, lease recovery, automatic reconnect or restart continuity is needed to pass.

A substantive spoken lookup remains R1a. VoiceClaw failure must preserve the healthy NemoClaw agent; automated recovery is outside this first proof.

## 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).

Blocked by:

- [#11747: [R0] Specify the v1-to-VoiceClaw onboarding handoff and scoped connection fixtures](https://github.com/NVIDIA/NemoClaw/issues/11747)
- [#11750: [R0] Verify one plain OpenClaw deployment on NemoClaw v1 through Brev CLI](https://github.com/NVIDIA/NemoClaw/issues/11750)
- [#11751: [R0] Apply voice integration intent on v1 and hand off to VoiceClaw onboarding](https://github.com/NVIDIA/NemoClaw/issues/11751)
- [#11749: [R0] Connect VoiceClaw to one v1 agent with a scoped credential](https://github.com/NVIDIA/NemoClaw/issues/11749)

## Category

Testing

Start by reading VOICECLAW_NEMOCLAW_API_CONTRACT.md and the blocked issues #11747, #11750, #11751, and #11749. Use the Brev CLI test steps as the entry point, then verify the acceptance checklist: a real client reaches the intended v1 agent through VoiceClaw with scoped authentication, protected handoff, readiness and failure checks, and repeatable redacted evidence.

Written by the indexing model from the issue text.