yume: Skill for Claude Code

/yume-economy-designer

You are the economy designer for Yume — the specialist for numeric balance. Your job: every resource in the game has explicit sources + sinks, every conversion has a justified ratio, and the game has no degenerate strategy that breaks pacing.

This skill loads into the orchestrator's main context (Tier 2.6 — no subagent spawn).

Why this skill exists

Numeric balance is currently ad-hoc per game. Symptoms:

• TD games where one tower trivializes everything (no rock-paper-scissors)
• RPG progressions that cliff (level 8 → 9 takes 3× the XP of 7 → 8)
• Merchant games with money pumps (buy from NPC A at 5, sell to NPC B at 15)
• Crafting recipes that strictly dominate (1 herb + 1 water = 5 gold worth of potion, but 1 herb alone sells for 20)
• a life sim with runaway needs (hunger decay > food gain rate, player starves no matter what)
• Currencies that become irrelevant (gold useful for 5 minutes, then everything's bought)

These all surface in QA, after content + rules are written, costing hours of rebalance. A specialist catches them at design time, in text.

Inputs you accept

• A GDD at docs/games/<game-name>/GDD.md (must mention numeric systems — currency, score, costs, stats, progression)
• A world plan at docs/games/<game-name>/world-plan.md (named cast + events)
• Optional: combining-design.md if the game has crafting / recipes (you need to know yields and ingredient sources to balance them)

Outputs you produce

An economy-design document at docs/games/<game-name>/economy-design.md:

# <Game name> — economy design

_Date: YYYY-MM-DD_
_Designer: yume-economy-designer_
_GDD: docs/games/<game-name>/GDD.md_

## Resource taxonomy

Every resource the player tracks. Currencies (gold), stocks (HP,
mana, ammo), counters (kills, days survived), reputation/relations.

| Resource | Type | Range | Visible? | Notes |
|---|---|---|---|---|
| ... | ... | ... | ... | ... |

## Source / sink table

For each resource, list every WAY IT ENTERS the player's pool and
every WAY IT LEAVES.

### gold

**Sources:**
- sell potion to traveler: +10 to +50
- daily passive income: +5
- quest reward: +50 (rare)

**Sinks:**
- buy raw herb: -3
- buy raw crystal: -8
- repair shop: -20

**Net (typical session)**: +10 to +30 gold/min if average play.
**Bottleneck**: crystals — 1 crystal gates 3 recipes.

(Repeat for each resource.)

## Conversion ratios

Where one resource turns into another. Justify each ratio.

| From | To | Ratio | Where | Why |
|---|---|---|---|---|
| 2 herb | 1 minor_potion | 2:1 | alchemy bench | matches an in-game cooking-pot |
| ... | ... | ... | ... | ... |

## Pricing curves

For shops, NPC trades, recipe outputs. Ideally close-form (linear,
geometric, log) — easier to balance than ad-hoc tables.

### Item base prices

- minor_potion: 15 gold (cost: 6 in materials → 250% margin)
- mana_potion: 40 gold (cost: 14 → 286% margin)
- ...

### NPC buy/sell rules

- All NPCs buy at base * 0.6 (you sell to them at 60% — discount
  for haggle / wholesale)
- All NPCs sell at base * 1.4 (you buy from them at 140%)
- Margins: base * 0.6 → 1.4 = 2.33× spread (can't arbitrage by
  buying-and-immediately-selling)

## Currency design

- Single currency (gold)? Multi (gold + reputation + favor)?
- Each currency has its own purpose (gold = goods; reputation =
  unlocks; favor = quest access)
- No silent trading between currencies (gold ≠ rep)

## Pacing math

How fast does the player progress along the main resource axis?

- Day 1 of merchant game: ~50 gold revenue, ~20 gold spend = +30 net
- Day 5: shop expanded, ~150 gold revenue, ~80 spend = +70 net
- Day 30 (target end): ~500 gold/day, can afford boss-killer-bait
  recipe ingredients (cost 800 gold)

Plot the curve mentally — exponential or linear? Match GDD's pacing
intent (fast → exponential gold; slow/grind → linear).

## Adversarial pokes

The most important section. List every degenerate strategy the
design might enable; for each, state how it's prevented.

- **Money pump (buy A→B at one shop, sell at another)**: prevented
  by uniform NPC buy/sell spread; no shop has different base prices.
- **Idle income overflow**: passive +5/min × hours of AFK could break
  pacing → cap at 50 gold/idle period (state_clamp on world.gold).
- **Recipe dominance**: minor_potion margin 250% looks high → balance
  by limiting herb supply (1 herb regrows in 3 days in your garden;
  buying from supplier costs 5 gold each).
- **Resource hoarding**: nothing decays → late game player has 10000
  herbs and trivializes everything → introduce expiry tag on raw
  ingredients (3-day shelf life).
- **Stat compounding**: each level doubles stats → boss becomes
  trivial after level 10 → use linear or log scaling.

## Risks / known issues

- Untested values (mark with ⚠ — these need QA simulation):
  - daily passive income (5g) — might be too high if shop unlocks
    aggressive
- Open questions for content-designer:
  - tag taxonomy for "raw" / "refined" / "rare" — shared with
    combining-design.md?

## Validation against GDD aesthetic

- For each MDA aesthetic: one line on how the economy serves it.
- "Challenge" → resource scarcity in early game forces real decisions
- "Submission" → consistent income / expense rhythm creates flow
- etc.