QA-PW-141 — Retries configured without a flake-triage loop
Generated from the rule registry. Get the same detail from the CLI with
mjolnir explain QA-PW-141.
| Field | Value |
|---|---|
| Severity | warning |
| Confidence | high |
| Tier | extended |
| Measured FP rate | 9% (n=33) |
| Evidence level | E1 |
| QA impact | False-green risk (FALSE-GREEN) |
| False-positive risk (author estimate) | low |
| Autofix available | no |
| Languages | typescript, javascript |
| Frameworks | playwright |
| Detection strategy | LEXICAL (regex heuristic) |
| Introduced in | v0.3.8 |
Why this fails in production
Retries convert intermittent failures into silent passes. Without a forensics/triage step consuming retry data, flaky tests pass forever and real regressions hide behind lucky reruns.
What gets flagged (real detector output)
retries: undefined with no visible flake-triage loop.Example from this rule's own must-fire fixture: tests/fixtures/QA-PW-141/must-fire/playwright.config.ts
The fix
Keep retries <= 2 and feed retry outcomes into the evidence path (explain --evidence), or add a reporter so flaky passes are reviewed.
Confirmed NOT to fire on the corresponding clean pattern
Verified against tests/fixtures/QA-PW-141/must-not-fire/playwright.config.ts — a legitimate, similar-looking pattern this rule correctly leaves alone.
Corpus-measured false-positive risk
Real occurrence counts from npm run corpus:audit against actively-maintained OSS repos — reproduce yourself, don't just trust this table (see the false-positive audit):
| Repo | Occurrences |
|---|---|
| calcom-cal | 2 |
| dubinc-dub | 1 |
| grafana-grafana | 1 |
| hashicorp-vault | 1 |
| negative-fixtures | 24 |
| nextauthjs-next-auth | 1 |
| positive-fixtures | 24 |
| vercel-next-js | 1 |
| vitest-dev-vitest | 1 |
| withastro-astro | 2 |