What problem does it solve? Landing page reviews suffer from self-assessment bias: the agent that built the page tends to mark its own work as passing, and existence checks (alt attributes, lang tags) pass even when text is unreadable on screen. This Skill separates aesthetic/layout checks into deterministic pixel-level measurement and routes subjective marketing judgment to independent reviewers who did not build the page. ## Core Features & Use Cases - Deterministic aesthetic measurement: The measure_lp.mjs script uses Playwright to render the page across viewports and color schemes, checking contrast ratios, type scales, tap targets, line length, heading orphans, first-viewport H1/CTA presence, and image cropping against documented thresholds, splitting results into disqualifying failures and warnings. - Independent marketing scoring: Three reviewers with distinct perspectives (first-time target reader, skeptical non-target, competitor comparison) score 10 rubric items from 0-4, with medians taken and rewrite suggestions required for low scores. - Gated workflow: Qualitative review only runs after deterministic checks pass, preventing wasted reviewer tokens on broken pages. - Use Case: Before publishing a landing page, run the measurement script to catch a light-mode headline with 1.1:1 contrast that existing validators missed, then have independent reviewers score whether the copy communicates the service within 3 seconds. ## Quick Start Ask the AI to evaluate your landing page by providing its URL, for example: "Review the landing page at https://example.com/lp before we publish it."
