gsd-verify-work

Conversational user acceptance testing is run by executing a single-test-at-a-time workflow that captures user responses and session gates to validate built features. The system maintains persistent session state and produces an auditable markdown report for each phase.

Yes, when UAT tests fail, the system automatically diagnoses gaps and drafts verified fix plans ready for execution. It prepares actionable fix plans that can be directly handed off to engineers for iterative issue resolution.

The best way to maintain persistent state is by emitting phase-specific UAT markdown reports, such as phase_{n}-UAT.md, that track all results, decisions, and next steps. This creates an auditable artifact for stakeholder handoffs and release validation.

Session gating in conversational QA works by presenting one test at a time in plain text and capturing user responses before proceeding. This ensures clarity and reproducibility by validating each test outcome individually before moving to the next step.

Yes, this workflow is designed for development and release candidates, meaning built features must exist to validate. It validates feature parity from an end-user perspective and generates diagnostic findings for any gaps discovered.

A conversational UAT workflow removes the friction of ad-hoc testing by structuring validation into single-test sessions with automated diagnosis. It captures outcomes systematically and generates phase-specific reports with ready-to-run fix plans for engineering handoff.