What

Closes the deferred Wave 5.5 architectural goal by collapsing `TicketWorkspaceRail`'s 38-prop flat interface into 8 named bundles + 2 primitives. Each bundle maps to one cohesive feature area of the rail:

Plus two flat primitives that don't cluster: `workspaceCatalogLoading` and `currentResponse`.

Why

Wave 5.5 deliberately narrow-scoped its decomposition. The plan called out the 26→~8 prop refactor as "a separate, larger PR outside Wave 5" — flagged in the post-Wave-5 verification audit, now addressed. Bundling delivers:

• Call-site grouping: each bundle is assembled once from a related set of shell state+handlers, making the DraftTab shell's workspacePanel builder read as 8 cohesive groups instead of a 38-line flat prop list.
• Future decomposition path: a follow-up can move each bundle into its own typed hook without changing the rail's prop surface again.
• Exported bundle interfaces (`TicketWorkspaceRailIntakeBundle`, etc.) can be consumed by future rail sub-components.

How

1. Define 8 exported bundle interfaces in `TicketWorkspaceRail.tsx`.
2. Rewrite the rail's internal destructuring + JSX to reference bundled values (`intake.data`, `similarCases.items`, `packs.handoffPack`, etc.).
3. Update `DraftTab.tsx`'s `workspacePanel` builder to construct the 8 bundles inline.
4. Update `TicketWorkspaceRail.test.tsx` fixture to build the bundled shape; adjust 3 assertion paths to the new names (`props.similarCases.onCompareLast`, `props.kits.onApply`, `props.runbooks.onCopyProgress`, `props.personalization.onChange`).

Testing

• `pnpm tsc --noEmit` — clean.
• `pnpm test` — 241 pass, unchanged count. Same three tests on the rail, same behavior coverage, just new prop shape.
• `pnpm ui:gate:static` — exit 0.

Risk / Notes

• No runtime behavior change. Same callbacks fire at the same moments with the same arguments.
• Only consumer is `DraftTab.tsx`; the shell is updated in the same PR so no external drift.
• The rail's public behavior (rendered DOM, click handlers, aria attributes) is byte-identical.