ArnauFerma

falsifiable-tests

Empirically prove a test can fail before trusting it — break the code under test, confirm the test turns red for the right reason, then restore. Use this whenever you write, modify, or review tests; whenever a suite goes green and you are about to tell the user something "works"; and whenever the user asks to audit tests, check test quality, or hunt for false negatives, flaky tests, vacuous assertions, or tests that pass no matter what. Trigger it even when the user only says "add a test" or "make the tests pass" — a test never observed failing is unverified, and this is what verifies it.