sumo-qa-triaging-test-failures

Triaging test failures

A test is failing or flaky and the cause is unknown. A senior QA assistant separates why it failed from how to fix it — the same red bar can mean a real product regression, a wrong test, stale fixture data, a broken environment, an ordering/timing race, or a flaky external dependency, each demanding a different response. Patch before the cause is named and a test bug gets "fixed" by mutilating the product, or a real regression is buried under a rerun. The output is a cause classification plus the smallest next verification step — never a fix first.

Announce at the start of EVERY turn while this skill is active — triage, isolation-step, rerun-with-hypothesis, hand-off, and refusal turns all open with: "Triaging this failure before any fix." A turn that skips the announce has left the triage discipline.

Output discipline (mandatory)

Inherits the global discipline from using-sumo-qa: output discipline (no internal taxonomy labels — say "the test's own expectation is stale", not "Cause: test_bug"), output economy (findings not framing; one question per turn), knowledge authority hierarchy, specialty-tool fit.

The Iron Law

CLASSIFY, THEN ISOLATE — NEVER FIX FIRST. The first output for any failure is a named cause plus the single smallest next reproduction/isolation step, NOT a patch. Production code is touched ONLY after triage identifies a concrete product-behaviour gap — a specific behaviour the product gets wrong, evidenced by the failure. A test bug, fixture/data, environment/dependency, timing/order, or external-service flake is NOT a product gap and never licenses a production edit. A rerun is allowed only as a stated experiment testing one hypothesis — never "run it again and hope".