Concrete pre-merge demo

Review AI-generated work before you trust it.

An agent generated a plausible onboarding feature. Round Table Workspace turns that confidence into a small review panel and a practical decision: ship, revise, or reject.

Star on GitHub Back to overview 
$ ./rtw ship-check "Should we merge this AI-generated onboarding feature?"

Decision: revise
Confidence: medium
Why:
- useful feature shape
- missing test coverage around persisted state
- unclear success state for first-time users
- public copy should avoid claiming a completed onboarding flow
Generated

Checklist

Welcome tasks and first-run guidance.

Generated

State

Completion state stored in local storage.

Generated

Copy

New empty dashboard language.

Question

Merge?

The change looks plausible, but needs review.

Round-table findings
ProductUseful direction, but the first-run path is still vague.
EngineeringNeeds tests around local-storage fallback and reset behavior.
RiskEmpty-state copy makes a support promise the app does not yet prove.
User advocateChecklist labels are clearer, but the success state is missing.
Decision

Revise before merge

The feature should not be rejected, but it should not ship as-is. The review turns a confident generated change into concrete follow-up work. 

Next:
1. add tests for disabled storage, reset, and completed states
2. add a visible success state after the checklist is done
3. weaken "guided onboarding" to "onboarding checklist"
4. rerun the project test suite before merge

What this demo proves

This is an illustrative transcript, not host-live or provider-live validation evidence. It demonstrates the repository's intended review pattern: make risks, missing evidence, and next actions visible before AI-generated work becomes trusted work.