Quality streak loop

Run realistic cases until N pass in a row, restarting the streak whenever a failure is found and fixed.

Evaluation Matthew Berman evaluation qa regression

Use this to force a workflow through varied examples without accepting a lucky pass.

Choose the required streak length before testing.
Generate or select realistic scenarios one at a time.
Run each scenario and record evidence.
When a scenario fails, document it, fix the root cause, and add regression coverage.
Reset the streak after every failure.
Continue until the target streak passes.

Stop when the latest N scenarios pass in a row and every earlier failure has a regression guard.