What problem does it solve? Writing reliable end-to-end browser tests is slow and error-prone: scripts often lack reproducibility, skip assertions, and produce no evidence when they fail. This Skill guides you through creating reviewable Playwright test scripts, running them deterministically, and automatically collecting video, trace, HAR, and screenshot evidence for every run. ## Core Features & Use Cases - Template-based test authoring: Start from ready-made pytest templates (auth, list, form, dashboard) and record interactions with playwright codegen, then add expect() assertions and page-role markers. - Reproducibility review checklist: Enforces checks for deterministic data, marker coverage, assertion completeness, and explicit wait strategies before any test runs. - Evidence collection by default: Every run captures video, Playwright traces, HAR network logs, and failure screenshots into a structured reports/<run-id>/ directory, with optional overlay captions and Google Drive upload. - Flexible browser connection: Run against local containerized Chromium or attach to a remote Chrome over CDP to reuse logged-in sessions, including SSH-based host Chrome launching. - Use Case: After implementing a new login flow, record the flow with codegen, turn it into a reproducible pytest with assertions, then run ./scenario-test/run.sh to get a video and trace proving the flow works. ## Quick Start Ask the AI to write a Playwright E2E test for your login page using the auth template and run it with video evidence enabled.
