openai/gpt-chat-latest → A (4:1)

Side A repairs broken end-to-end OAuth test infrastructure by fixing query parsing (`str/split` with regex), reading POST bodies correctly (`getRequestBody`), handling null tokens/states safely, correcting redirect responses, and wrapping mock handlers to avoid crashes, along with improving Playwright DOM selection and synchronization. Side B adds useful CLI visibility by formatting and displaying existing connectivity statistics with tests, but it is primarily a presentation enhancement rather than a correctness fix restoring critical test functionality.