NVIDIA / NVIDIA/NemoClaw

refactor(onboard): consolidate desired-state authorities

Open
#7,701 0 comments 0 reactions 1 assignee Claimed by @cv View on GitHub
area: architecture area: onboarding refactor
Dominant language
TypeScript
Stars
22.5k
Forks
3.1k
Avg merge
1d 1h
Merged PRs (30d)
715

Description

## Ownership

@cv owns this issue.

Do not start, assign, or submit this work unless @cv transfers ownership in this issue.

## Problem

Onboard, rebuild, resume, and channel mutation can derive the same desired state from different sources.

The lifecycle map identifies split ownership for messaging intent, target resolution, conflict policy, and policy persistence.

## Desired behavior

- The registry-persisted messaging plan is authoritative for an existing sandbox.
- One target-resolution policy defines when current input can replace recorded intent.
- One declarative conflict policy serves onboard, rebuild, and channel mutation.
- One policy-persistence owner reconciles create-time, live, and channel changes.
- A persisted plan stores logical bindings only. Apply-time code binds credentials and live capabilities.

## Change boundary

Consolidate one decision authority in each PR.

Keep interactive prompt policy at the caller when prompt behavior differs.

Do not define a public plugin or lifecycle SDK.

## Security

Persist no credential values.

Reject a target or binding conflict before an external or destructive effect.

## Validation

- Test source precedence for onboard, resume, rebuild, and channel mutation.
- Test conflict parity across callers.
- Test registry and session reconciliation after a successful apply.
- Test failure before mutation when recorded and current intent conflict.

## Completion

Each listed desired-state decision has one owner and one tested precedence policy.

Contributor guide

Open the contributing guide

Assessment

This issue has not been assessed yet.

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.