The QA Suite Kit — the exact prompt that made Claude Code write a passing Playwright suite in 2m56s for $0.78, and the sabotage protocol that proves your suite has teeth

The companion kit to the video. We pointed Claude Code at a real store app with zero tests: one prompt, 2 minutes 56 seconds, $0.78 — 14 passing Playwright tests, green on the first run, with role-based selectors and its own webServer config. Then we planted 5 one-line bugs to catch it lying. It caught 4 instantly. The one it missed was the money bug: a discount shown on the checkout page while the customer was charged full price — because no test ever asserted the amount actually charged. One added prompt line ('follow the money all the way through') rebuilt the suite to 15 tests and caught all five. INSIDE (6 pages): (1) THE PROMPT, VERBATIM — every line annotated, including the run-it-yourself instruction that turns code generation into a QA engineer, plus the follow-the-money line. (2) THE CLAUDE.MD TEMPLATE — the 8 lines that made the run take 3 minutes instead of an hour. (3) THE 5-STEP SYSTEM — file, prompt, the one permission setting, sabotage check, CI. (4) THE SABOTAGE PROTOCOL — 5 one-line mutations to plant in your own app, what each one proves, and the anatomy of the money miss ($23.38 promised vs $25.98 charged). (5) THE RECEIPTS — every stage with wall clock and cost, plus honest limits. Independent kit from Hyperautomation Labs; not affiliated with Anthropic or Microsoft (Playwright).

By submitting, you'll be subscribed to the Hyperautomation AI Report newsletter — one practical AI email per week. Unsubscribe anytime.

Free. No spam. Unsubscribe anytime.