What problem does it solve? Test suites often contain hidden quality problems such as hard waits, missing assertions, shared state, and flakiness patterns that pass CI but erode reliability. This Skill audits existing test files against a fixed criteria registry and produces a deterministic 0-100 quality score with severity-ranked findings and recommended fixes. ## Core Features & Use Cases - Rubric-Based Test Audit: Evaluates test files against pinned criteria (disabled tests, hard waits, determinism, isolation, fixtures, data factories, network-first patterns) with fixed severities, so two reviewers reach the same verdict. - Framework-Aware Analysis: Adapts to Playwright, Jest, Cypress, Vitest, Pact contract tests, and Maestro mobile flows, loading only the knowledge fragments relevant to the detected stack. - Scored Review Report: Generates a structured Markdown report with a quality score, letter grade, computed recommendation (Approve / Request Changes / Block), code-level fix examples, and optional inline TODO comments in the reviewed files. - Use Case: Before merging a pull request, run the review on the changed test files to catch a committed test.only or a race condition between navigation and network interception, and receive a report the team can act on. ## Quick Start Ask the agent to review the test quality of the files in your tests directory and generate a scored review report.
