# task-qa-sweep-3c7b8b3d - the QA sweep runner read 31 Playwright failures as
# `status=passed`: the parser walked one suite level and every spec sits under
# a `test.describe`, and `failed` was the parser's count, not Playwright's.
# This PR is the one that makes the nested-suite assertions pass.
tests/test_qa_agent_runner.py
