Fixture-backed local transcript

See the review gate in one minute.

One question goes in. A small panel reviews product value, engineering readiness, risk, and user clarity. The output is a practical ship, revise, or reject decision.

Screenshot-ready one-minute demo

Before one confident AI answer becomes trusted work, run a local review gate.

$ ./rtw ship-check "Should we merge this AI-generated feature?"

Decision: revise
Confidence: medium

Panel:
- product: revise
- engineering: ship
- risk: revise
- user-advocate: revise
Decision: revise Ship after evidence and claim wording improve.
ProductRevise: value is still too vague.
EngineeringShip: small enough after tests pass.
RiskRevise: claim boundary needs proof.
User advocateRevise: show a concrete example.
Fixture-backed local transcript. No host-live or provider-live claim. github.com/MarkDonish/round-table-workspace

Summary

Useful enough to continue, but revise positioning, evidence, and user-facing examples before claiming it is launch-ready.

revise

Product

The user value should be stated as an observable outcome before shipping.

ship

Engineering

The change is small enough to ship when tests and local validation pass.

revise

Risk

Claim boundaries and rollback notes should be explicit before launch copy is promoted.

revise

User advocate

The public README should show a concrete before/after example, not only architecture language.

It catches

Unsupported host-live or provider-live claims before they become public launch copy.

It names

Missing evidence, weak positioning, and the exact next step instead of vague approval.

It preserves

A local evidence trail that can be committed, audited, and rerun before merge.

Next:
1. Run ./rtw doctor --quick and the unit test suite.
2. Add one concrete demo transcript or screenshot to the README.
3. Keep public claims local-first unless host-live/provider-live evidence exists.