framework-quality-auditlisted

# Framework Quality Audit ## Purpose EITR's core product bet (see `docs/architecture/README.md` Section 1, "Introduction & Goals") is that the generated framework is the deterministic first layer of a larger SDET ecosystem, not a disposable scaffold. That means generation quality has to be checked periodically, per stack combination, with a report that is **comparable across runs** — same categories, same scoring rigor, same layout — so drift and regressions show up as a diff in scores over time, not as a differently-shaped essay each time. This skill is intentionally a **process + report-shape** definition, not a scorecard. It carries no memorized scores or verdicts from any prior run — each invocation is a fresh, unbiased read of whatever the generator currently produces for the requested stack. Never seed a new audit with a previous run's numbers. ## When to use Invoke this skill whenever the user asks for a quality/maturity audit, review, or scorecard of an EITR-generated framework — for the TS+Playwright combo or any other supported language + automation-tool pairing. If the user names a stack, audit that stack. If they don't, ask which stack (or default to whichever sample project already exists in the repo, if any). ## The 19 fixed categories (exact names, exact order, exact grouping) Never rename, reorder, merge, split, or add/remove a category between runs — that is what makes scores comparable across audits. If a category is genuinely inapplicable to a given