PF1e (SRD spells/acid-arrow.md): "Range long (400 ft. + 40 ft./level)" - the JSON declared medium. Fixed to long. The spell's ongoing acid (2d4 per extra round, one per three caster levels, unless neutralized) is not modeled: the engine has no per-spell damage-over-time subsystem, documented in README rather than built.
Ultraworked with [Sisyphus](https://github.com/code-yeongyu/oh-my-openagent)
Co-authored-by: Sisyphus <clio-agent@sisyphuslabs.ai>
PF1e (SRD spells/burning-hands.md): "1d4 points of fire damage per caster level (maximum 5d4)" — the JSON encoded 1d4+1, a flat bonus the spell never grants. Base is now 1d4. Per-caster-level scaling and the 15-ft. cone are not modeled by the engine (no CL-scaled dice, no cone template); documented in README rather than built.
Ultraworked with [Sisyphus](https://github.com/code-yeongyu/oh-my-openagent)
Co-authored-by: Sisyphus <clio-agent@sisyphuslabs.ai>
PF1e (SRD spells/bleed.md): "Saving Throw Will negates" and the target takes 1 point of damage per round. Two divergences: the JSON declared a Fortitude save, and the bleeding condition never set the engine's bleed_damage field, so the spell was a no-op. Bleed now saves Will; _apply_condition_effect sets bleed_damage = 1 when a spell applies bleeding, and _decrement_durations lands the final tick as the last bleeding condition expires before clearing the field (heal-path clearing already existed in _stop_bleeding). README updated.
Ultraworked with [Sisyphus](https://github.com/code-yeongyu/oh-my-openagent)
Co-authored-by: Sisyphus <clio-agent@sisyphuslabs.ai>
PF1e (SRD spells/sound-burst.md): "Every creature in the area takes 1d8 points of sonic damage and must succeed on a Fortitude save to avoid being stunned for 1 round" — the Fortitude partial negates the stun only, never the dice. The damage component declared half_on_save: true, halving the sonic damage on a successful save. DamageEffect gains a save_negates flag (default True, existing behavior unchanged); sound_burst.json declares half_on_save: false + save_negates: false so the 1d8 is full regardless while the condition keeps save_negates. README updated.
Ultraworked with [Sisyphus](https://github.com/code-yeongyu/oh-my-openagent)
Co-authored-by: Sisyphus <clio-agent@sisyphuslabs.ai>
PF1e CRB 'Big and Little Creatures in Combat': a creature's natural
reach is a disk (a Large creature still threatens and strikes adjacent
squares), while a reach weapon doubles the wielder's natural reach but
'can't strike at their natural reach or less' (a ring whose dead zone is
the natural reach). The engine keyed the ring off reach_ft > 5, so
reach-10 natural attacks (giant belostomatid bite/claw, mounted
hooves) wrongly got a 5-ft dead zone.
AttackSpec gains reach_weapon (parsed from the combatant JSON); the
disk-vs-ring rule is centralized in _melee_reach_zone_cells and applied
consistently to weapon selection, melee usability, threatened squares
and charge end placement. Esha's Lance is the only reach weapon in the
shipped data; Large natural attackers keep the default (disk).
Ultraworked with [Sisyphus](https://github.com/code-yeongyu/oh-my-openagent)
Co-authored-by: Sisyphus <clio-agent@sisyphuslabs.ai>
PF1e CRB (Magic - Ranged Touch Attacks): spells delivered by ray require a
ranged touch attack (BAB + Dex vs touch AC, no range increment) to hit a
hostile target; a miss wastes the slot. SpellSpec gains requires_ranged_touch,
set on the 8 SRD-confirmed rays (Ray of Frost, Ray of Enfeeblement, Scorching
Ray, Acid Arrow, Acid Splash, Disrupt Undead, Shard of Chaos, Spear of Purity).
Magic Missile and area spells are unaffected. Simplification (documented): no
size modifier / Weapon Finesse on the roll.
Ultraworked with [Sisyphus](https://github.com/code-yeongyu/oh-my-openagent)
Co-authored-by: Sisyphus <clio-agent@sisyphuslabs.ai>
SpellSpec gains a cast_time field (seven budget codes; the JSON loader
also accepts RAW phrasings like "1 round" or "1 minute"). The pure
budget filters take a CastTimeResolver so a cast_spell occupies the slot
its casting time maps to: standard->standard, swift->swift (competing
with deed/inspire courage and dropped by a consumed immediate),
immediate/free->free, full_round and one_round->full_round (one_round
clamped: resolved end-of-turn, documented approximation), long->dropped
before the budget filter with the new downgraded:cast_time reason. The
engine binds its registry lookup via _cast_time_of (unknown names fall
back to standard, keeping transcripts byte-identical);
_choose_spell_action never selects long casts; the staggered filter now
drops full-round casts and spell_combat via the effective slot
(_FULL_ROUND_KINDS becomes dead and is deleted). Summon Monster I/II are
the only data spells marked 1 round. Metamagic full-round casting for
spontaneous casters is out of scope (future hook documented in
_cast_time_of).
Ultraworked with [Sisyphus](https://github.com/code-yeongyu/oh-my-openagent)
Co-authored-by: Sisyphus <clio-agent@sisyphuslabs.ai>
Rename monster.py → combatant.py with _KNOWN_FEATS lookup table
that resolves feat name strings to AbilitySpec constants (Power Attack,
Iron Will, Toughness, Weapon Focus, etc.). Unknown feats produce
minimal AbilitySpec (data preserved for future implementation).
Migrate all 8 PC JSONs (data/pcs/) with complete feat/spell lists
extracted from Foundry sheets — including traits, racial abilities,
and all spells. Monsters (goblin, orc) unchanged (no feats/spells).
Update all imports, add backward-compatible aliases (load_monster,
MonsterLoadError), update CLI to use load_combatant.
Gate: 377 tests, ruff clean, basedpyright clean.
All 8 sheets from fiches_personnages/ converted to the strict Combatant
JSON schema (data/monsters convention) so the CLI can run PCs directly:
--side players data/pcs/esha.json
- id derived from the file stem; source field points at the original sheet
- oni: placeholder attacks without encoded damage ('une Pioche', 'Weapon')
left out, documented in notes (pick stats ambiguous light/heavy in the
rules KB; no identification possible for 'Weapon')
- README: PC section now documents data/pcs usage
- tests: parametrized load check + golden spots for the 8 PC files
ctan