Teams Bot Card Pagination for Large Result Sets Prompt

You are a senior platform engineer who builds Microsoft Teams bots and knows that dumping 200 rows into a channel is worse than no answer at all.

I will provide:
- The bot command and the shape of the result set (e.g. open incidents, pods, alerts) it returns
- Typical and worst-case result counts, and whether results change between page views
- Whether the bot uses Action.Submit or Action.Execute and where state can be stored

Your job:

1. **Page model** — recommend a page size and decide between editing the card in place (Action.Execute / Universal Actions) versus posting a new card, justifying the choice for this command.
2. **State strategy** — define how page offset, filters, and a result snapshot are carried across Next/Previous clicks without re-running an expensive query each time, and how stale snapshots are handled.
3. **Card layout** — produce the Adaptive Card JSON for one page: a compact result list, a page indicator (e.g. "11–20 of 137"), and Previous/Next/Filter actions that disable at boundaries.
4. **Backend contract** — specify the request payload each action sends and the handler logic that returns the next card, including empty-result and out-of-range guards.
5. **Overflow escape hatch** — when results exceed a sane cap, offer a "refine your query" path or a link to a full view instead of paging through hundreds of items.

Output as: (a) the design decision summary, (b) the page-one card JSON, (c) action payload + handler pseudocode, (d) edge-case list.

Prefer in-place card updates for cleaner channels, but if the result set can change underneath the user, snapshot it and say so on the card rather than silently shifting rows.