S4 visual-verify method — capture a stable baseline, deterministically re-capture the candidate under identical conditions, diff within tolerance with dynamic regions masked, and require a human/view_image classification of every diff. Use at the S4 acceptance gate of web-app / game-production (any UI/render change that must be judged intended movement vs accidental drift).
$npx -y skills add mtarcure/claude-vibe-squad --skill visual-regression-baselineInstalls into the current project.
Run `npx skills use "https://github.com/mtarcure/claude-vibe-squad" --skill "mtarcure/claude-vibe-squad/visual-regression-baseline"` and follow the generated skill instructions now. Read its complete output, redirecting it to a temporary file first if necessary. Resolve relative paths from the supporting-files directory it provides.
Use the skills in "https://github.com/mtarcure/claude-vibe-squad" that are relevant to the current task. Run `npx skills add "https://github.com/mtarcure/claude-vibe-squad"` and select the relevant skills, then follow their instructions.