What problem does it solve? Taking a project with no end-to-end tests to a maintained Playwright suite normally requires ad-hoc decisions about scaffolding, journey discovery, coverage, and reporting. This Skill defines a gated eight-phase workflow so an agent (or human) can run the full bootstrap deterministically, with enforced exit criteria between phases. ## Core Features & Use Cases - Eight-phase pipeline: scaffold, groundwork, happy-path, journey-mapping, coverage-expansion, bug-discovery, secrets-sweep, and report, each with explicit exit criteria and per-deliverable commits. - State-machine enforcement: a status ledger at tests/e2e/docs/onboarding-status.json plus workflow-reviewer subagents gate every phase, pass, and cycle transition, blocking silent phase skips. - Dual run modes: interactive use inside Claude Code or hands-off execution from an external automated CLI driver, with standard vs depth strictness levels. - Use Case: Point the agent at a running web app with no tests; it scaffolds Playwright config, maps user journeys, writes specs per priority tier, runs adversarial bug hunts, sweeps secrets into .env, and produces a QA summary deck. ## Quick Start Ask the agent to onboard this project and automate your app's URL from zero, then choose standard or depth run mode when prompted.
