[Feature request] Split and refine the Mobile accessibility checklist

### Description

The first version of the Mobile accessibility checklist is now live in the [repository](https://github.com/github/annotation-toolkit/blob/main/checklists/mobile-checklist.md) and on the [primer.style docs site](https://primer.style/accessibility/tools-and-resources/checklists/mobile-checklist/) (shipped via #41, closing #36).

In practice, the mobile checklist ended up about twice as long as the existing web checklists. That length makes it harder to scan and harder to hand to a single audience without a lot of "skip this section" caveats. This issue tracks the follow-up edits and improvements needed to bring it in line with the rest of the toolkit.

### Requested change

Refactor the Mobile checklist into multiple, more focused variants, mirroring how the web checklists are split today:

- **Split by audience:** Separate the content into a mobile **designer** checklist and a mobile **engineer** checklist, so each role only sees items relevant to their phase of work.
- **Split by depth:** Offer a **full** checklist (the comprehensive list we have now) and a **high-impact** short list that surfaces the most critical items for teams who can't run through the full thing on every project.
- **Rebuild the Figma component:** Update the Checkpoints component in the Figma library so the new variants are selectable as properties, consistent with the existing web checklist component. This is a heavier lift than the markdown work since the current

Comece com checklists/mobile-checklist.md e compare-o com as checklists web existentes para definir as variantes designer, engineer, full e high-impact. Em seguida, inspecione o componente Checkpoints na biblioteca do Figma e o trabalho entregue em #41; considera-se concluído quando as variantes Markdown puderem ser usadas sem ressalvas sobre pular seções e o componente Figma expuser as novas variantes como propriedades selecionáveis.

Escrita pelo modelo de indexação a partir do texto da issue.