mockup-fidelity

{
  "name": "mockup-fidelity",
  "description": "Validate that an implemented React or React Native UI faithfully reproduces a reference mockup, then update the code to close every gap, by measuring the rendered tree rather than eyeballing it or reading source. Treats the mock as the source of truth, inventories every frame including the modals and states the named screens open, and inverts the burden of proof so a difference is a defect until an external citation proves it intentional. Its distinguishing feature is that it refuses to certify a property its browser engine cannot actually measure: a runtime preflight sets each declaration through a real stylesheet rule and reads it back with two different sentinels, and any detector class that fails is reported as inconclusive with the engine's own reason string instead of quietly agreeing. On the sanctioned engine nine classes fail that probe, including box-shadow and placeholder colour, the two properties most often silently wrong. Diffs structure before styling, pairs text-less icons and dividers by bounding box, crops and pixel-diffs every paired element, and emits a present/divergent/absent ledger, a functional-gaps document, and a three-valued exit code where inconclusive is neither a pass nor a failure. Carries a second measurement engine for native macOS, Electron and web-view targets, and for the classes a browser engine reports nothing for: it drives the proctor skill to read the accessibility tree, the compositor's resolved layer values and frame trustworthiness, establishing that engine's own capability tier before a style finding may claim anything. Use it when someone asks whether a build matches its design, to pixel-match a screen, to audit UI fidelity after a port or migration, or to find what a mock contains that was never built.",
  "version": "3.4.2",
  "author": {
    "name": "Fledgeling",
    "email": "[email protected]"
  }
}