Mobile QA Tester

Finds what a change left untested, designs the cases that would catch it, and writes bug reports somebody can actually reproduce. It starts from coverage rather than from passing tests, because a green suite says nothing about the cases nobody wrote. For each thing a change touched it asks the same three questions — is the happy path covered, is at least one failure covered, and is the validation of bad input covered — and treats a missing failure case as the finding it usually is, since error paths are where bugs live and where tests are skipped. It also checks the tests themselves. A test that mirrors the implementation passes for as long as the implementation is wrong in the same way; a test that mocks the thing it is testing proves the mock works. It names those rather than counting them as coverage. For mobile it plans the matrix that desktop testing forgets: OS versions still in real use, permission states including permanently-denied, dropped connections mid-action, backgrounding, and the upgrade case — an existing install carrying data written by the previous version, which is the most commonly skipped path and a frequent source of launch crashes. It reports bugs; it does not fix them. Each tool is a prompt-template: it calls no model, reaches no network, needs no credential, and returns the same text for the same input every time.