write-e2e

What problem does it solve? Writing meaningful end-to-end tests requires discovering routes, ranking user flows by risk, and asserting real causality rather than incidental page state. This Skill automates that pipeline: it analyzes a web application's codebase, scores user flows by criticality, generates Playwright specs that assert observable outcomes, and honestly reports what was actually proven versus merely written. ## Core Features & Use Cases - Route and flow discovery with criticality scoring: Scans routes, forms, API endpoints, and auth patterns, then scores each candidate flow on five weighted signals (mutation, auth, sensitivity, traffic, coverage) with confidence labels. - Causality-first spec generation: Every scenario fills a causality contract (trigger, decisive event, pre-state, post-state, visible oracle, cleanup) before any spec code is written, enforced by ten critical quality gates E2E-Q1 through E2E-Q10. - Verification ladder with honest states: Each flow earns exactly one state — GENERATED, STATIC_CHECKED, VERIFIED_LOCAL, VALIDATED_LIVE, BLOCKED, or FAILED — so unexecuted specs are never reported as passing coverage. - Origin safety and consent gates: Classifies every target as LOCAL, STAGING, or EXTERNAL_UNKNOWN with fail-closed network mocking, blocking mutations against non-local origins without explicit consent flags. - Use Case: Point it at a Next.js app with --auto to generate the three highest-scored Playwright specs, execute them locally, and receive a coverage registry showing exactly which flows were verified. ## Quick Start Ask the AI to generate Playwright E2E tests for the most critical user flows in this project and run them locally.