What problem does it solve? After converting a design prototype into a built web page (e.g., an EDS/AEM site), subtle regressions slip through: stretched images, dropped CTAs, headings rendered in the wrong slot, or fonts silently falling back. Manual eyeballing misses these, and a single pixel-diff tool cannot see structural content errors. ## Core Features & Use Cases - Pixel/Layout Probe: Detects stretched images, dropped max-width wraps, blank renders, surface/ground color flips, and image-count gaps via computed-style metrics and screenshots. - Structural Content + Typography Probe: Extracts an ordered, role-classified inventory (headings, eyebrows, CTAs, body) from each page and diffs them to catch missing content, role swaps, and rendered-font forks using a width probe. - Stack-Agnostic Profiles: Ships with eds and generic profiles so the same engines work for AEM/EDS conversions or any prototype-vs-build fidelity check. - Use Case: After running a prototype-to-EDS conversion, run both probes against the prototype URL and the branch preview URL, fix every red structural flag, and re-run until the pass bar is met. ## Quick Start Ask the AI to run the diff skill comparing your prototype URL against the built page URL with the eds profile and report any layout or content mismatches.