task_games__cards__longest_run_length

Contract

1. Domain: games
2. Scene id: cards
3. Public task id: task_games__cards__longest_run_length
4. Supported query_id values: single
5. Answer schema: integer_value
6. Annotation schema: bbox_set
7. Program schema: max(lengths(consecutive_rank_runs(cards_in_display_order))); scene=cards; scope=longest_run_length

Program Contract

Program: max(lengths(consecutive_rank_runs(cards_in_display_order))); scene=cards; scope=longest_run_length

Candidate set: the visible game board, pieces, tokens, cards, tiles, marked state, legal-move cues, result panels, and labeled options inside the longest_run_length objective scope. Operands: visible scene state and prompt-bound operands named by lengths, consecutive_rank_runs, cards_in_display_order, cards, longest_run_length. Operation: evaluate max over the candidate set using the visible game state, rules, legal moves, comparisons, counts, simulations, or option-selection constraints encoded in the program expression; generation enforces a unique final answer. Output binding: answer uses the integer schema; generation binds a unique final answer. Annotation witnesses: annotation uses the bbox_set schema; the prompt/annotation contract defines the minimal visual witnesses. Query ids: single.

Reasoning Operations

Families: counting, ranking, topology

Generation Notes

1. Prompt wording comes from src/trace_tasks/resources/prompts/games/cards/games_cards_v1.json.
2. Annotation is projected from the same generated game state used for answer verification.