feat(combat): add full attack action with BAB iterative attacks

- New Action kind 'full_attack': manufactured weapons (count=1) gain BAB
  iteratives at +6/+11/+16 (-5/-10/-15 from attack bonus, capped at 4);
  natural weapons (count>1) delegate to _attack (no iteratives).
- Extracted _resolve_swing from _attack: one swing, log, stats, target-down
  check. Both _attack and _full_attack use it. AttackResult gains base_bonus
  field so the transcript shows the actual bonus used per iterative.
- default_policy now returns full_attack instead of attack when in range.
  Non-breaking for BAB < 6 (single swing, identical transcript).
- Refactored round_no to self._current_round instance variable, removing it
  from 6 method signatures and fixing PLR0913 on _resolve_swing.
- 4 new tests: BAB+6 two iteratives, BAB+11 three iteratives, target drops
  mid-full-attack, BAB<6 single swing. README and docstring updated; test
  count 202 -> 206.

README.md

@@ -278,7 +278,10 @@ Règles modélisées :
 - Attaques multiples par action : une arme avec `count` > 1 (« 2x Talons »)
   résout `count` balayages indépendants dans la même action ; les balayages
   restants sont perdus si la cible tombe (inconsciente ou morte) en cours de
-  rafale. Les itératifs de BAB (2e attaque à BAB +6…) ne sont pas modélisés.
+  rafale. Attaque à outrance (action à round complet) : à BAB +6/+11/+16, des
+  attaques supplémentaires à -5/-10/-15 ; les armes naturelles (`count` > 1)
+  ne gagnent pas d'itératifs. Plusieurs armes naturelles en une seule attaque
+  à outrance ne sont pas modélisées.
 - Mêlée : allonge d'une case ; mouvement : un déplacement parcourt jusqu'à la
   vitesse du combattant en cases le long du plus court chemin réel (champ de
   coût Dijkstra depuis la cible, diagonales 5-10-5), en s'arrêtant adjacent à
@@ -306,9 +309,8 @@ appliquées dans la résolution) :
 - Sorts, jets de sauvegarde, conditions et états.
 - Attaques d'opportunité, manœuvres de combat.
 - Effets mécaniques de hauteur/élévation.
-- Tailles Large+ (2×2), allonge > 5 ft, itératifs de BAB, actions spéciales
-  (attaque à outrance, charge, pas de placement, retraite) et actions
-  immédiates hors-tour.
+- Tailles Large+ (2×2), allonge > 5 ft, actions spéciales
+  (charge, pas de placement, retraite) et actions immédiates hors-tour.
 
 ## Architecture
 
@@ -342,7 +344,7 @@ appliquées dans la résolution) :
 La gate de validation complète (tests + lint + types) :
 
 ```bash
-uv run pytest -q          # 202 tests
+uv run pytest -q          # 206 tests
 uv run ruff check src tests
 uv run basedpyright src   # mode strict
 ```

src/pf1e_simulator/combat.py

@@ -13,7 +13,11 @@ Phase 0 documented deviations from PF1e (conventions):
   (AoO, flanking, full attack, charge, withdraw) are deferred.
 - A weapon with `count` > 1 ("2x Talons") resolves `count` independent attacks
   in one attack action; remaining swings are lost once the target drops
-  (down or dead) mid-routine. BAB iterative attacks are not modeled.
+  (down or dead) mid-routine. BAB iterative attacks (full-round action
+  ``full_attack``): at BAB +6/+11/+16, additional attacks at -5/-10/-15 from
+  the weapon's attack bonus; iteratives are only computed for weapons with
+  ``count == 1`` (natural multi-attacks do not gain iteratives). Multiple
+  natural weapons in a single full attack are not modeled.
 - Natural 1 always misses; natural 20 always hits and threatens a crit.
 - A confirmed crit multiplies the total damage (dice + flat bonus) by crit_mult.
 - DR applies once, after crit multiplication; any bypassing type defeats DR.
@@ -60,6 +64,8 @@ _COVER_AC_BONUS = 4  # PF1e: partial cover grants +4 AC
 _RANGE_INCREMENT_PENALTY = 2  # PF1e: -2 per full range increment beyond the first
 _MAX_RANGE_INCREMENTS = 10  # PF1e: projectile weapons shoot up to 10 increments
 _FLANK_BONUS = 2  # PF1e: +2 melee attack when ally threatens opposite side
+_ITERATIVE_PENALTY = 5  # PF1e: each BAB iterative is at -5 from the previous
+_MAX_ITERATIVES = 4  # PF1e: BAB +16 gives 4 attacks (at +16, +11, +6, +1)
 
 _STEP_DELTAS: tuple[Pos, ...] = (
     (-1, -1),
@@ -117,6 +123,7 @@ class AttackResult:
     ac: int
     penalty: int = 0
     flank_bonus: int = 0
+    base_bonus: int = 0
 
 
 @dataclass(frozen=True)
@@ -124,14 +131,17 @@ class Action:
     """What a policy wants a combatant to do this turn.
 
     Kinds map to the six PF1e action types: ``attack`` is a standard action
-    (single attack), ``move`` is a move action, ``full_round`` consumes the
-    whole turn (full attack, charge, withdraw — deferred), ``swift`` and
-    ``free`` are minor actions, ``immediate`` is an off-turn reaction
-    (deferred). ``wait`` is a no-op. The engine executes a policy-returned
-    sequence per turn; the default policy currently emits one action.
+    (single attack), ``move`` is a move action, ``full_attack`` is a full-round
+    action (BAB iteratives), ``full_round`` is a generic full-round action
+    (charge, withdraw — deferred), ``swift`` and ``free`` are minor actions,
+    ``immediate`` is an off-turn reaction (deferred). ``wait`` is a no-op.
+    The engine executes a policy-returned sequence per turn.
     """
 
-    kind: Literal["attack", "move", "wait", "swift", "free", "immediate", "full_round"]
+    kind: Literal[
+        "attack", "move", "wait", "swift", "free", "immediate", "full_round",
+        "full_attack",
+    ]
     target_id: str | None = None
 
 
@@ -172,19 +182,20 @@ class CombatEngine:
         self._policy = policy if policy is not None else default_policy
         self._transcript: list[str] = []
         self._stats: dict[str, _LiveStats] = {s.combatant.id: _LiveStats() for s in states}
+        self._current_round = 0
 
-    def _take_turn(self, state: CombatantState, round_no: int) -> bool:
+    def _take_turn(self, state: CombatantState) -> bool:
         """Execute one combatant's full turn; return True if the battle is over."""
         actions = self._policy(self, state)
         before = len(self._transcript)
         for action in actions:
             if not state.active:
                 break
-            self._execute(state, action, round_no)
+            self._execute(state, action)
             if self._winner_side() is not None:
                 break
         if state.active and len(self._transcript) == before:
-            self._log(f"round {round_no} {state.combatant.id}: wait")
+            self._log(f"round {self._current_round} {state.combatant.id}: wait")
         return self._winner_side() is not None
 
     def run(self) -> CombatResult:
@@ -192,10 +203,11 @@ class CombatEngine:
         order = self._roll_initiative()
         round_no = 1
         while round_no <= self._round_cap:
+            self._current_round = round_no
             for state in order:
                 if not state.active:
                     continue
-                if self._take_turn(state, round_no):
+                if self._take_turn(state):
                     break
             if self._winner_side() is not None:
                 break
@@ -265,14 +277,20 @@ class CombatEngine:
         return 0
 
     def resolve_attack(
-        self, attacker: CombatantState, target: CombatantState, weapon: AttackSpec
+        self,
+        attacker: CombatantState,
+        target: CombatantState,
+        weapon: AttackSpec,
+        *,
+        bonus_override: int | None = None,
     ) -> AttackResult:
         """Roll one attack (with crit confirm and damage) and apply it."""
         roll = self._rng.d20()
         dist_ft = self._grid.distance(attacker.pos, target.pos) * _SQUARE_FT
         penalty = self.range_penalty(weapon, dist_ft)
         flank = self._flanking_bonus(attacker, target) if weapon.kind == "melee" else 0
-        total = roll + weapon.attack_bonus + penalty + flank
+        base_bonus = bonus_override if bonus_override is not None else weapon.attack_bonus
+        total = roll + base_bonus + penalty + flank
         ac = target.combatant.ac.total
         occupied = frozenset(
             s.pos for s in self._states if s.active and s is not attacker and s is not target
@@ -291,7 +309,7 @@ class CombatEngine:
         if hit:
             if roll != _NATURAL_ONE and roll >= weapon.crit_range:
                 confirm = self._rng.d20()
-                confirm_total = confirm + weapon.attack_bonus + penalty + flank
+                confirm_total = confirm + base_bonus + penalty + flank
                 crit = confirm != _NATURAL_ONE and confirm_total >= ac
             damage = sum(c.formula.roll(self._rng) for c in weapon.damage) + weapon.damage_bonus
             if crit:
@@ -300,7 +318,7 @@ class CombatEngine:
         target.hp -= damage
         return AttackResult(
             hit=hit, crit=crit, damage=damage, roll=roll, total=total, ac=ac,
-            penalty=penalty, flank_bonus=flank,
+            penalty=penalty, flank_bonus=flank, base_bonus=base_bonus,
         )
 
     def _apply_dr(self, damage: int, weapon: AttackSpec, target: CombatantState) -> int:
@@ -340,63 +358,105 @@ class CombatEngine:
             return active_sides.pop()
         return None
 
-    def _execute(self, state: CombatantState, action: Action, round_no: int) -> None:
+    def _execute(self, state: CombatantState, action: Action) -> None:
         target = next((s for s in self._states if s.combatant.id == action.target_id), None)
-        if action.kind == "attack":
+        if action.kind in ("attack", "full_attack"):
             if target is None or not target.active:
                 return
             weapon = self.weapon_for(state, target)
             if weapon is None:
                 return
-            self._attack(state, target, weapon, round_no)
+            if action.kind == "full_attack":
+                self._full_attack(state, target, weapon)
+            else:
+                self._attack(state, target, weapon)
         elif action.kind == "move":
             if target is None:
                 return
-            self._move(state, target, round_no)
+            self._move(state, target)
         elif action.kind == "wait":
-            self._log(f"round {round_no} {state.combatant.id}: wait")
+            self._log(f"round {self._current_round} {state.combatant.id}: wait")
 
-    def _attack(
-        self, attacker: CombatantState, target: CombatantState, weapon: AttackSpec, round_no: int
-    ) -> None:
+    def _resolve_swing(
+        self,
+        attacker: CombatantState,
+        target: CombatantState,
+        weapon: AttackSpec,
+        *,
+        label: str,
+        bonus_override: int | None = None,
+    ) -> bool:
+        """Resolve one swing, log it, update stats. Return True if target dropped."""
         live = self._stats[attacker.combatant.id]
         target_live = self._stats[target.combatant.id]
+        hp_before = target.hp
+        result = self.resolve_attack(attacker, target, weapon, bonus_override=bonus_override)
+        live.hits += int(result.hit)
+        live.crits += int(result.crit)
+        live.damage_dealt += result.damage
+        target_live.damage_taken += result.damage
+        outcome = "CRIT" if result.crit else "HIT" if result.hit else "MISS"
+        flank_str = f"+{result.flank_bonus}(flank)" if result.flank_bonus else ""
+        bonus_str = (
+            f"{result.base_bonus}{result.penalty:+d}{flank_str}"
+            if result.penalty
+            else f"{result.base_bonus}{flank_str}"
+        )
+        line = (
+            f"round {self._current_round} {attacker.combatant.id}: {label} "
+            f"vs {target.combatant.id} "
+            f"d20={result.roll}+{bonus_str}={result.total} "
+            f"AC {result.ac} -> {outcome}"
+        )
+        if result.hit:
+            line += f" {result.damage} damage ({hp_before}->{target.hp})"
+        self._log(line)
+        if target.dead:
+            self._log(
+                f"round {self._current_round} {attacker.combatant.id}: "
+                f"{target.combatant.id} dead"
+            )
+            return True
+        if not target.active:
+            self._log(
+                f"round {self._current_round} {attacker.combatant.id}: "
+                f"{target.combatant.id} down"
+            )
+            return True
+        return False
+
+    def _attack(
+        self, attacker: CombatantState, target: CombatantState, weapon: AttackSpec
+    ) -> None:
         for swing in range(1, weapon.count + 1):
-            hp_before = target.hp
-            result = self.resolve_attack(attacker, target, weapon)
-            live.hits += int(result.hit)
-            live.crits += int(result.crit)
-            live.damage_dealt += result.damage
-            target_live.damage_taken += result.damage
-            outcome = "CRIT" if result.crit else "HIT" if result.hit else "MISS"
             label = weapon.name if weapon.count == 1 else f"{weapon.name} #{swing}"
-            flank_str = f"+{result.flank_bonus}(flank)" if result.flank_bonus else ""
-            bonus = (
-                f"{weapon.attack_bonus}{result.penalty:+d}{flank_str}"
-                if result.penalty
-                else f"{weapon.attack_bonus}{flank_str}"
-            )
-            line = (
-                f"round {round_no} {attacker.combatant.id}: {label} vs {target.combatant.id} "
-                f"d20={result.roll}+{bonus}={result.total} "
-                f"AC {result.ac} -> {outcome}"
-            )
-            if result.hit:
-                line += f" {result.damage} damage ({hp_before}->{target.hp})"
-            self._log(line)
-            if target.dead:
-                self._log(f"round {round_no} {attacker.combatant.id}: {target.combatant.id} dead")
-                return
-            if not target.active:
-                self._log(f"round {round_no} {attacker.combatant.id}: {target.combatant.id} down")
+            if self._resolve_swing(attacker, target, weapon, label=label):
                 return
 
-    def _move(self, state: CombatantState, target: CombatantState, round_no: int) -> None:
+    def _full_attack(
+        self, attacker: CombatantState, target: CombatantState, weapon: AttackSpec
+    ) -> None:
+        if weapon.count > 1:
+            self._attack(attacker, target, weapon)
+            return
+        n = min(
+            _MAX_ITERATIVES,
+            1 + max(0, (attacker.combatant.bab - 1) // _ITERATIVE_PENALTY),
+        )
+        for i in range(n):
+            label = weapon.name if n == 1 else f"{weapon.name} #{i + 1}"
+            bonus = weapon.attack_bonus - _ITERATIVE_PENALTY * i
+            if self._resolve_swing(
+                attacker, target, weapon, label=label, bonus_override=bonus
+            ):
+                return
+
+    def _move(self, state: CombatantState, target: CombatantState) -> None:
         path = self._move_path(state, target)
         if not path:
             return
         coords = "->".join(f"({p[0]},{p[1]})" for p in (state.pos, *path))
-        self._log(f"round {round_no} {state.combatant.id}: move {coords}")
+        self._log(f"round {self._current_round} {state.combatant.id}: move {coords}")
         state.pos = path[-1]
 
     def _move_path(self, state: CombatantState, target: CombatantState) -> list[Pos]:
@@ -443,13 +503,13 @@ Policy = Callable[[CombatEngine, CombatantState], tuple[Action, ...]]
 
 
 def default_policy(engine: CombatEngine, state: CombatantState) -> tuple[Action, ...]:
-    """Attack the nearest enemy in range; otherwise approach at full speed then try to attack."""
+    """Full-attack the nearest enemy in range; otherwise approach then full-attack."""
     target = engine.nearest_enemy(state)
     if target is None:
         return (Action(kind="wait"),)
     if engine.weapon_for(state, target) is not None:
-        return (Action(kind="attack", target_id=target.combatant.id),)
+        return (Action(kind="full_attack", target_id=target.combatant.id),)
     return (
         Action(kind="move", target_id=target.combatant.id),
-        Action(kind="attack", target_id=target.combatant.id),
+        Action(kind="full_attack", target_id=target.combatant.id),
     )

tests/test_combat.py

@@ -865,3 +865,87 @@ def test_flanking_requires_ally_on_same_side() -> None:
         "round 1 tgt: short sword vs atk d20=5+2=7 AC 13 -> MISS",
         "battle over: draw after 1 rounds",
     )
+
+
+# --------------------------------------------------------------------------- #
+# Full attack (CRB: BAB iteratives at BAB, BAB-5, BAB-10, BAB-15)             #
+# --------------------------------------------------------------------------- #
+
+
+def test_full_attack_bab6_two_iteratives() -> None:
+    """BAB +6 gives two attacks: at +6 and +1."""
+    war = make_combatant("war", attack_bonus=6, hp=10, speed=0)
+    tgt = make_combatant("tgt", hp=20, ac=15, attack_bonus=0, speed=0)
+    states = [
+        make_state(war, "players", (3, 4)),
+        make_state(tgt, "monsters", (4, 4)),
+    ]
+    engine = make_engine([10, 8, 12, 3, 5, 2], states, round_cap=1)
+    result = engine.run()
+    assert result.transcript == (
+        "initiative: war d20=10+0=10",
+        "initiative: tgt d20=8+0=8",
+        "round 1 war: short sword #1 vs tgt d20=12+6=18 AC 15 -> HIT 3 damage (20->17)",
+        "round 1 war: short sword #2 vs tgt d20=5+1=6 AC 15 -> MISS",
+        "round 1 tgt: short sword vs war d20=2+0=2 AC 13 -> MISS",
+        "battle over: draw after 1 rounds",
+    )
+
+
+def test_full_attack_bab11_three_iteratives() -> None:
+    """BAB +11 gives three attacks: at +11, +6, +1."""
+    war = make_combatant("war", attack_bonus=11, hp=10, speed=0)
+    tgt = make_combatant("tgt", hp=30, ac=15, attack_bonus=0, speed=0)
+    states = [
+        make_state(war, "players", (3, 4)),
+        make_state(tgt, "monsters", (4, 4)),
+    ]
+    engine = make_engine([10, 8, 12, 3, 10, 2, 5, 2], states, round_cap=1)
+    result = engine.run()
+    assert result.transcript == (
+        "initiative: war d20=10+0=10",
+        "initiative: tgt d20=8+0=8",
+        "round 1 war: short sword #1 vs tgt d20=12+11=23 AC 15 -> HIT 3 damage (30->27)",
+        "round 1 war: short sword #2 vs tgt d20=10+6=16 AC 15 -> HIT 2 damage (27->25)",
+        "round 1 war: short sword #3 vs tgt d20=5+1=6 AC 15 -> MISS",
+        "round 1 tgt: short sword vs war d20=2+0=2 AC 13 -> MISS",
+        "battle over: draw after 1 rounds",
+    )
+
+
+def test_full_attack_stops_on_target_down() -> None:
+    """Remaining iteratives are lost when the target drops mid-full-attack."""
+    war = make_combatant("war", attack_bonus=6, hp=10, speed=0)
+    tgt = make_combatant("tgt", hp=5, ac=15, attack_bonus=0, speed=0)
+    states = [
+        make_state(war, "players", (3, 4)),
+        make_state(tgt, "monsters", (4, 4)),
+    ]
+    engine = make_engine([10, 8, 12, 5], states, round_cap=1)
+    result = engine.run()
+    assert result.transcript == (
+        "initiative: war d20=10+0=10",
+        "initiative: tgt d20=8+0=8",
+        "round 1 war: short sword #1 vs tgt d20=12+6=18 AC 15 -> HIT 5 damage (5->0)",
+        "round 1 war: tgt down",
+        "battle over: players win in 1 rounds",
+    )
+
+
+def test_full_attack_low_bab_single_swing() -> None:
+    """BAB < 6 gives one attack — same as a regular attack, no '#1' label."""
+    war = make_combatant("war", attack_bonus=3, speed=0)
+    tgt = make_combatant("tgt", hp=10, ac=15, attack_bonus=0, speed=0)
+    states = [
+        make_state(war, "players", (3, 4)),
+        make_state(tgt, "monsters", (4, 4)),
+    ]
+    engine = make_engine([10, 8, 12, 2, 5], states, round_cap=1)
+    result = engine.run()
+    assert result.transcript == (
+        "initiative: war d20=10+0=10",
+        "initiative: tgt d20=8+0=8",
+        "round 1 war: short sword vs tgt d20=12+3=15 AC 15 -> HIT 2 damage (10->8)",
+        "round 1 tgt: short sword vs war d20=5+0=5 AC 13 -> MISS",
+        "battle over: draw after 1 rounds",
+    )