ts-pw-test-assistant

ts-pw-test-assistant is an artifact-driven, multi-agent AI Quality Engineering system built on Playwright + TypeScript. It is described as a practical implementation platform for an agentic AI for Quality Engineering series. The system demonstrates how specialized AI agents work with an existing Playwright + TypeScript automation framework to transform a JIRA requirement into structured test scenarios, acceptance-criteria coverage, framework-aware Playwright tests, semantic and technical quality validation, live browser execution, execution evidence, a consolidated QE report, and release-gated Git changes. The AI layer does not replace the Playwright automation framework; it operates on top of the framework and dynamically discovers and reuses existing Page Objects, fixtures, locators, helpers, assertions, and conventions. Agents communicate through persisted artifacts, and the workflow does not depend on transient LLM conversation memory to pass important state between stages. The workflow is controlled by explicit quality gates and stop conditions rather than treating a successful AI response or a green Playwright exit code as sufficient evidence. A critical rule is that the AI must not invent Page Object methods, selectors, or framework capabilities that do not exist; if a required capability cannot be found, the workflow produces a framework_gap_report.md and stops.