> ## Documentation Index
> Fetch the complete documentation index at: https://docs.etonecarg.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Prompt Library

> Short prompt examples and the smallest safe Sports MCP workflows that usually answer them well.

# Prompt Library

These examples are intentionally short and product-shaped.

| User prompt                                  | Recommended path                                                                                                                |
| -------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------- |
| “What is on tonight?”                        | `list_watchable_schedule` and then `list_live_slate` if the UX needs now-playing context                                        |
| “Catch me up on this game”                   | `resolve_entities` only if needed, then `list_schedule` and `get_event_summary` or `get_event_center`                           |
| “Show me the current league picture”         | `resolve_entities` and then `get_competition_hub`                                                                               |
| “Tell me about this player”                  | `resolve_entities` and then `get_participant_profile`                                                                           |
| “Where can I watch this match?”              | identify the right event and then `get_watch_availability`                                                                      |
| “What else should I watch?”                  | `resolve_entities` or `get_program_context`, then `list_watchable_schedule` or `list_programs_for_entity`, and rank in the host |
| “How does this season compare to last year?” | call `get_competition_hub` multiple times and compare in the host                                                               |

## Use the dedicated prompt pages when

* the prompt can branch into more than one source family
* the workflow depends on whether IDs are already known
* the product needs low-token follow-on steps after the first answer