/quality-engineering-test-healing

name: quality-engineering-test-healing
description: Failure taxonomy and allowed/forbidden repairs for a failing E2E test. Use when a Playwright/Maestro/Detox/XCUITest/Espresso/Appium test fails and you must decide whether to repair the test or route to a real bug.
metadata:
  triggers:
    files:
      - "test-results/**"
      - "playwright-report/**"
    keywords:
      - heal test
      - failing e2e
      - selector repair
      - test healer
      - fix the test
      - timed out waiting for

Quality Engineering: Test Healing

**Priority: P0 (CRITICAL)**

Failure Classes

`SELECTOR_DRIFT` (id/locator changed) · `TIMING_SYNC` (race/wait, no state change) · `DATA_ENV` (fixture/seed/env stale) · `INFRA` (network/runner/flake) · `REAL_REGRESSION` (product behavior actually changed).

Classify from evidence: trace/screenshot/DOM diff supports a drift/timing/data explanation, or the product diff shows an intentional behavior change (REAL_REGRESSION).

Allowed Repairs

Move the locator up the selector ladder (e.g., `getByRole` if the ID changed); replace a sleep with an explicit state wait; fix a stale fixture/seed; retry only for `INFRA`.

Forbidden Repairs

Never weaken an assertion. Unchanged assertion text does NOT authorize weakening fixtures, state preconditions, or locator specificity (e.g., replacing targeted test seeds with generic/empty defaults or widening selectors to match unintended elements). Never add `test.skip`/`fixme`; a test that must stop gating goes to quarantine per `quality-engineering-flaky-triage` with a ticket + expiry, and keeps running. Never widen a matcher. Never inflate a timeout by more than 2x. Never blind `--update-snapshots`. Never catch-and-continue. Never touch production code — that is `REAL_REGRESSION`, not a heal.

Verdicts

`HEALED` (repair verified by 3 consecutive green reruns, ASSERTION_DELTA: none, fixture and selector integrity preserved) · `REAL_BUG_DO_NOT_HEAL` (product behavior changed or real defect exposed; route to dev-fix) · `QUARANTINE_CANDIDATE` (intermittent failure across isolated reruns; route to flaky-triage) · `BLOCKED` (no evidence artifact, or no stable locator target: route to `specialist-testid-inserter`).

Red Flags

"the assertion is unchanged so loosening the fixture is fine" · "broaden selector so any element matches" · "the assertion was too strict anyway" · "the product changed so update the expected value" · "just add retries so it goes green" — do not heal with these rationalizations. All are `REAL_BUG_DO_NOT_HEAL` or `QUARANTINE_CANDIDATE` in disguise, never `HEALED`.

References

• [Failure Taxonomy Signals](references/failure-taxonomy.md)
• [Repair Catalog](references/repair-catalog.md)
• [Forbidden Repairs](references/forbidden-repairs.md)
• [Verdict Contract](references/verdict-contract.md)
• [Flaky Triage](../quality-engineering-flaky-triage/SKILL.md) — quarantine contract for `QUARANTINE_CANDIDATE`
• [Visual Baseline](../quality-engineering-visual-baseline/SKILL.md) — screenshot failures before `REAL_REGRESSION`