Fix flaky visual tests

A flaky diff comes back changed when nothing actually changed. The screenshot is not lying - it faithfully caught something that moves between runs, like the clock or an animation frame. Remove that source and the flake goes with it.

• The clock - relative timestamps like "2 hours ago", or a date rendered from Date.now().
• Animations - an infinite CSS or JS animation caught mid-frame at a different point each run.
• Live data - a component that fetches a real API renders different content every time.
• Randomness - Math.random, generated ids, shuffled orders.
• Third-party widgets - maps, embeds, ad slots that load their own moving content.
• Web fonts - text captured before the font swaps in reflows on the next run.

• Freeze the clock to a fixed instant so relative-time labels are identical every run.
• Disable animations with animation: none under prefers-reduced-motion, which UI Verify emulates at capture - a near-zero animation-duration (or animation-iteration-count: 1) leaves an infinite animation running and can still flake under render load.
• For a JavaScript animation no CSS media query can reach - a charting library's animate-on-mount prop, a <canvas> loop - branch on `isUIVerify()` to render the final frame only while capturing (isAnimationActive={!isUIVerify()}), including at SSR/build time via a UI_VERIFY env var.
• Feed components fixture data instead of a live fetch.
• Seed or stub any randomness to a fixed value - and when the value is chosen during server rendering, like a random image picked in a React component, seeding the browser RNG cannot reach it, so branch on `isUIVerify()` instead.
• Stub third-party embeds behind a static placeholder.
• Wait for fonts to load before the screenshot is taken.
• Wait for the loaded content before the screenshot, not a loading skeleton or spinner - a capture taken mid-fetch diffs against the loaded baseline.

A test that navigates and then asserts a URL can take its screenshot while the page is still fetching, so the capture is a spinner or a skeleton that diffs against the loaded baseline. Wait for the real content to appear before the test ends - a data row, not just its container. One trap on a page that swaps content in: a generic selector can match stale content still in the DOM mid-transition and resolve too early, so prefer waiting for network-idle or for an element that only exists once the new content has rendered.

For Storybook, control the variation inside the story - fixed props, frozen time, fixture data. For real pages driven by Playwright or Vitest, archive-replay bakes the page in at record time, so a live API or a feature flag cannot shift it on replay - but you still freeze the clock and stub moving widgets before you record. Animations are the exception to baking: the archived CSS re-runs on replay, so the same prefers-reduced-motion freeze applies to archives too, not just Storybook.