VladyslavDmitriiev

Multi-agent QA pipeline: plain-language feature specs become Playwright tests through a writer/judge loop on different models, a real run, human review, and a mutation gate. Zero LLM calls in CI.