Show Card set in Game State: card_reward

[judgeou]

It will be useful

[Gennadiyev]

What's that, bundles? I thought card bundles were supported via #9 ...

Got the profile or run file for that scenario?

[judgeou]

These are the cards currently held by the player; I believe this information helps the AI ​​make correct decisions.

# Game State: card_reward

**Act 1** | Floor 2 | Ascension 0

## Player (You)
**静默猎手** — HP: 70/70 | Gold: 114

### Relics
- **蛇之戒指**: 在每场战斗开始时，额外抽2张牌。
- **精准剪刀**: 拾起时，从你的牌组中移除1张牌。

## Card Reward Selection
Choose a card to add to your deck:

- [0] **斗篷与匕首** (1 energy) [Skill] Common — 获得6点格挡。 将1张小刀添加到你的手牌。
- [1] **杂技** (1 energy) [Skill] Common — 抽3张牌。 丢弃1张牌。
- [2] **尖啸** (1 energy) [Skill] Common — 所有敌人在本回合失去6点力量。 消耗。

**Can skip:** Yes

## Keyword Glossary
- **格挡**: 在下个回合前，阻挡伤害。
- **力量**: 力量会增加攻击牌造成的伤害。
- **消耗**: 在战斗结束前移除。
- **小刀**: 造成4点伤害。 消耗。

The card rewards are currently not being displayed when making a selection.

[Gennadiyev]

Card rewards usually happen after a battle, and I expect that the model can rely on previous states to understand what is the current state of the deck. That said, a better solution can be to add a get deck method that stays up.

[CompleteDotTech]

Coverage note from the current submitted PR set: this should be addressed by #85.

The endpoint-audit PR includes full player data on every screen, including deck/deck_count, so the model has the current card set while in state_type: card_reward. The card reward choices themselves remain exposed under card_reward.cards.

I would keep this open until #85 is merged and the card_reward scenario is live-verified.