Clean is a starting point.

The source assigns twenty five points to message clarity, twenty to mobile ease, twenty to trust proof, fifteen to performance and twenty to conversion path. The weights total one hundred. Bars use four pixels per point to compare message clarity one hundred pixels with performance sixty pixels. These are manual review assertions, not a live page crawl, performance test, conversion analytics or verified client readiness. Project type changes a stored label, not these weights. A selected sum of fifty five says Needs a focused pass, while sixty says Promising foundation. Only one hundred says Client ready. Bars use seven pixels per point for fifty five and sixty, three hundred eighty five and four hundred twenty pixels. Every score is a weighted checkbox sum in multiples of five. The labels are source heuristic phrases and do not certify the site, verify that checks are true or predict visitor behavior. When all five checks are missing, the display orders them by descending weight: twenty five, three twenties, then fifteen. The JSON priority list retains original definition order and places performance before conversion path. Bars count five checks and five exported priorities at eighty pixels per item. The same names are present but their order differs. A checkbox change can also show a saved status mentioning the previous score because its handler closes over that old score. Both original behaviors are preserved. Export includes full checks and a physical timestamp, not live analytics.