Fright Path, Night 20: The Final Run.
Halloween night: every monster is out, and so is the team. Scheduled Runs keep watch hour after hour, the alert lands in Slack with the AI Summary following in its thread, and the next Run has already started.
does.qa #FrightPath
Trick or Tested.
Our logo dressed up for Halloween. Your tests don't have to be scary.
does.qa
#FrightPath
Fright Path, Night 19: I See Dead Tests.
Skipped, commented out, checking that true is true, and the report stays green. In DoesQA, a Flow that checks nothing is labelled No checks, and Pack Health scores your test pack out of ten.
does.qa #FrightPath
A failed Run only helps if someone sees it. Send passed and failed Runs to email, a webhook, Slack or Teams, and let the failure come to you.
docs.does.qa/configuration/n… #TestAutomation #QA
Fright Path, Night 18: My Agent Is a Friendly Ghost.
New commits overnight, and nobody at the keyboard: it's your coding agent. Connect it to DoesQA over MCP or the CLI, and it runs real journeys in real browsers, fixes what broke and hands you the Run as proof.
does.qa #FrightPath
Fright Path, Night 17: The Flatline.
The purchases chart drops to zero, but orders are still coming in: the purchase event had stopped being sent. Check tag manager events inside the journey, and a missing one fails the pull request's check before production.
does.qa #FrightPath
Some tools quietly self-heal while the selector drifts. DoesQA takes the other path: shared Elements, real changes a person confirms with a Run, and you change it once.
does.qa/blog/how-resilient-a… #QA
Fright Path, Night 16: The Button Nobody Could Hear.
To a screen reader, the basket says button, button, button, and every test still passes. With DoesQA, accessibility checks are steps in your Flows, and a button with no accessible name fails the test.
does.qa #FrightPath
Every test was green. The AI was rewriting the tests so they'd pass. Re-run with the original selectors, the pass rate was 14%.
The green tick is a signal, not a verdict.
does.qa/blog/every-test-was-… #QA
Fright Path, Night 15: The Order That Never Existed.
The thank-you page showed an order number, but nothing was saved or charged, and the test passed because it stopped there. Store the order number, call your own API with it, and check what was really saved.
does.qa #FrightPath
Run your test pack filtered by tags, override Values for just that Run, then save it as a Run Recipe so it's one click away next time. Run History keeps every Run.
docs.does.qa/runs/start-a-ru… #TestAutomation #QA
Playwright is free to start. Then 20 runners become 200, and you're maintaining runners, CI, reporting and debugging. The real cost is everything you build on top.
does.qa/blog/the-real-cost-o… #Playwright #QA
Fright Path, Night 14: The Hour That Happened Twice.
Last night UK clocks went back, and half past one happened twice: two customers, one table. One Flow, three Open steps on London, New York and Sydney time, and the midnight bug shows up in a working-day Run.
does.qa #FrightPath
Fright Path, Night 13: The Vampire Has No Reflection.
The Add to basket button passes every check, but a style change made it white on white. Vision steps check how the page actually looks: against a control image, or what a person should see, in plain English.
does.qa #FrightPath
Fright Path, Night 12: The Doppelgänger.
A feature flag sends some customers to a new checkout your tests never saw. Set the cookie or local storage value your flag reads, branch for both twins, and the new checkout fails at Pay now.
does.qa #FrightPath
Four test smells: the just-rerun-it test, DOM archaeology, the mock museum, and the green build nobody believes. Good tests fade into the background.
does.qa/blog/i-can-smell-you… #QA #FlakyTests
Your app can break on a day nobody deployed. Scheduled Runs keep checking: pick what to run, drag across the hours, and hear about failures in Slack, Teams, email or a webhook.
docs.does.qa/configuration/s… #TestAutomation #QA