Coverage Summary for Class: TheftModel (core.pipeline)

 // FILE: core/src/main/kotlin/core/pipeline/TheftModel.kt
 package core.pipeline
 
 import core.BalanceSettings
 import core.primitives.SalvagePolicy
 import core.rng.Rng
 import core.state.BoardContract
 import core.state.Hero
 
 /**
  * Theft decision model.
  *
  * ## Semantic Ownership
  * Answers exactly one question: **Why theft suspected?**
  *
  * ## Stability Gradient
  * STABLE: Pure decision logic with explicit rules.
  *
  * ## Determinism
  * - Uses exactly one RNG roll per decision.
  * - All inputs are explicit; no hidden state.
  *
  * ## Boundary Rules
  * - Must NOT change trophies stock.
  * - Must NOT emit events.
  * - Must NOT mutate state.
  */
 object TheftModel {
 
     /**
      * Decides whether theft occurred and computes reported trophies.
      *
      * @param hero The hero under suspicion.
      * @param board The contract's board terms.
      * @param baseTrophiesCount Trophies before potential theft.
      * @param rng Deterministic RNG for the roll.
      * @return [TheftDecision] containing theft flag and reported trophies.
      */
     fun decide(
         hero: Hero?,
         board: BoardContract?,
         baseTrophiesCount: Int,
         rng: Rng
     ): TheftDecision {
         if (hero == null || board == null || baseTrophiesCount <= 0) {
             return TheftDecision(
                 theftOccurred = false,
                 reportedTrophiesCount = baseTrophiesCount,
                 expectedTrophiesCount = baseTrophiesCount
             )
         }
 
         val theftChance = computeTheftChance(hero, board)
         val theftRoll = rng.nextInt(BalanceSettings.PERCENT_ROLL_MAX)
 
         return if (theftRoll < theftChance) {
             val stolenAmount = (baseTrophiesCount + 1) / 2
             val reportedAmount = (baseTrophiesCount - stolenAmount).coerceAtLeast(0)
             TheftDecision(
                 theftOccurred = true,
                 reportedTrophiesCount = reportedAmount,
                 expectedTrophiesCount = baseTrophiesCount
             )
         } else {
             TheftDecision(
                 theftOccurred = false,
                 reportedTrophiesCount = baseTrophiesCount,
                 expectedTrophiesCount = baseTrophiesCount
             )
         }
     }
 
     /**
      * Theft probability policy.
      *
      * Produces a single integer percentage in a fixed range.
      * All inputs are explicit. No hidden state is consulted.
      */
     private fun computeTheftChance(hero: Hero, board: BoardContract): Int {
         return when {
             board.salvage == SalvagePolicy.GUILD && board.fee == 0 ->
                 hero.traits.greed
             board.salvage == SalvagePolicy.GUILD && board.fee > 0 ->
                 (hero.traits.greed - board.fee / 2).coerceAtLeast(0)
             board.salvage == SalvagePolicy.HERO -> 0
             board.salvage == SalvagePolicy.SPLIT ->
                 ((hero.traits.greed - hero.traits.honesty) / 2).coerceAtLeast(0)
             else -> 0
         }
     }
 }
 
 /**
  * Decision output from theft resolution.
  *
  * This DTO is RNG-free and can be tested in isolation.
  */
 data class TheftDecision(
     /** Whether theft was detected. */
     val theftOccurred: Boolean,
     /** Trophies count after potential theft. */
     val reportedTrophiesCount: Int,
     /** Expected trophies count before theft. */
     val expectedTrophiesCount: Int
 )