What problem does it solve? Manually verifying that a pull request's UI changes actually work in a browser is slow and often skipped, letting visual regressions slip through. This Skill automates that QA pass: it boots the app, drives the changed surfaces through a configured browser provider, and produces screenshot-backed evidence without ever modifying source code. ## Core Features & Use Cases - Browser-driven UI verification: Derives a scoped QA scenario from the PR diff, exercises the affected routes through the configured browser provider, and captures a screenshot at each checkpoint. - Review-first gating and PR integration: Ensures the PR has been reviewed (invoking the review skill when not), claims the PR with a lock, and posts evidence inline as a PR comment with optional self-QA sign-off labels. - Local tracker-less mode: Runs against the current worktree without a tracker, writing machine-readable report.json and human-readable report.md artifacts for human review. - Use Case: After opening a PR that changes the login page, run the Skill against that PR number to get a posted comment containing step-by-step PASS/FAIL results and inline screenshots proving the new flow renders and behaves correctly. ## Quick Start Ask the agent to QA PR 1234's UI changes in a real browser and post screenshot evidence to the pull request.
